Incremental prognostic value of baseline 18F-FDG PET/CT tumour asphericity in breast carcinoma: association with de novo metastatic disease and survival outcomes

Incremental prognostic value of baseline 18F-FDG PET/CT tumour asphericity in breast carcinoma: association with de novo metastatic disease and survival outcomes | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Incremental prognostic value of baseline 18 F-FDG PET/CT tumour asphericity in breast carcinoma: association with de novo metastatic disease and survival outcomes Nitin Gupta This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9221004/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Tumour asphericity (ASP), derived from baseline 18F-FDG PET/CT, may capture morphologic heterogeneity linked to aggressive biology and adverse outcomes. Aim To evaluate the incremental value of baseline ASP for de novo metastatic disease at presentation and survival outcomes in breast carcinoma beyond clinical stage, receptor-derived subtype, Ki-67 and metabolic tumour burden. Methods Patient-level data (n = 180) were analysed. De novo metastasis (stage IV) was modelled using multivariable logistic regression with 5-fold cross-validated ROC analysis; model calibration was assessed using observed versus predicted risk. PFS and OS were analysed using Kaplan-Meier, log-rank tests and multivariable Cox proportional hazards regression. Results De novo metastatic disease was present in 37/180 patients (20.6%). In multivariable logistic regression, ASP (per 0.1 increase) independently predicted baseline metastasis (OR 3.60, 95% CI 2.02 to 6.43, p < 0.001) along with ln(TLG) (natural logarithm transformed TLG; OR 2.47, 95% CI 1.32 to 4.63, p = 0.0047). Out-of-fold AUC for metastasis prediction was 0.80 for ASP alone and 0.82 for the full model including ASP; calibration was satisfactory (calibration slope 1.00; Brier score 0.117). Higher ASP was also associated with treatment non-response (OR 1.99 per 0.1, p = 0.00033). On unadjusted analysis over full follow-up (not administratively censored), ASP > 0.30 (vs ASP ≤ 0.30) was associated with shorter PFS (median 24.7 vs 9.6 months; log-rank p = 0.0017) and OS (median 28.2 vs 10.8 months; log-rank p = 0.00011). Conclusion Baseline tumour asphericity derived from 18 F-FDG PET/CT is a simple, interpretable geometric biomarker that outperforms conventional metabolic indices and biomarker variables for baseline metastasis discrimination and is associated with treatment non-response and shorter median PFS and OS on unadjusted survival analysis. Prospective multicentre validation is warranted. Nuclear Medicine & Medical Imaging 18F-FDG PET/CT breast cancer tumour asphericity metabolic tumour volume total lesion glycolysis metastasis progression free survival overall survival Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction Breast cancer remains the most commonly diagnosed malignancy worldwide and a leading cause of cancer related mortality [ 1 ] . Although major therapeutic advances have improved outcomes, distant metastasis and treatment resistance continue to drive breast cancer deaths [ 2 ] . Baseline staging is therefore essential to guide treatment selection and prognostication, with whole body 18 F‑FDG PET/CT increasingly used in clinically high risk patients to detect extra axillary nodal and distant metastases and to establish a quantitative baseline for response assessment [ 3 , 4 ] . Beyond detection of overt metastatic disease, baseline PET/CT metrics may help contextualize tumour biology across molecular subtypes and inform risk-adapted systemic and locoregional strategies [ 2 , 4 ] . Robust quantification depends on harmonized acquisition, reconstruction and reporting as emphasized by contemporary procedure guidelines [ 5 ] . Conventional PET biomarkers such as SUVmax, metabolic tumour volume (MTV) and total lesion glycolysis (TLG) reflect uptake intensity and tumour burden; meta-analytic evidence supports their prognostic association in breast cancer, but clinically meaningful overlap persists among patients with similar clinicopathologic features [ 6 ] . Treatment response frameworks including the EORTC recommendations and PERCIST summarize response largely through uptake based measures [ 7 , 8 ] . Radiomics expands quantitative imaging by extracting descriptors of intensity, texture and geometry that may capture tumour phenotype and microenvironment [ 9 , 10 , 11 ] . and standardization initiatives such as the Image Biomarker Standardisation Initiative (IBSI) provide consensus definitions to improve reproducibility [ 12 ] . Among radiomic feature families, shape metrics are intuitive and comparatively less dependent on intensity normalization; asphericity (ASP = 1 - sphericity) summarizes deviation from a sphere and has shown independent prognostic value in other FDG avid cancers beyond conventional metabolic indices [ 13 , 14 ] . In breast cancer, PET derived radiomic indices including texture features have been linked to outcome, but the incremental value of primary tumour asphericity in routine clinical cohorts remains incompletely defined [ 15 ] . Because ASP is derived from surface area and volume, it is readily computable from standard tumour VOIs and is potentially amenable to routine reporting when robust segmentation is available. This study aimed to evaluate the prognostic and predictive utility of baseline 18 F-FDG PET/CT tumour asphericity (ASP) for baseline distant metastasis, progression free survival (PFS) and overall survival (OS) in breast carcinoma, and to determine whether ASP provides incremental value beyond clinical stage, molecular receptor status (ER/PR/HER2), Ki‑67 and standard metabolic tumour burden metrics (SUVmax, MTV and TLG). We also compared the discriminative performance of ASP against metabolic and biomarker variables for metastasis prediction and assessed its contribution in multivariable survival models. Why this matters clinically Early risk stratification is central to treatment selection and counselling in breast carcinoma [ 2 ] . Patients presenting with occult metastatic disease often require systemic therapy rather than curative local approaches, and those at high risk of early progression may benefit from intensified therapy and closer surveillance. Baseline distant staging investigations are generally reserved for higher risk presentations [ 3 ] and 18 F-FDG PET/CT can upstage disease and inform management in intermediate and high risk cohort [ 4 ] . If a simple PET derived morphologic metric such as ASP can help identify patients more likely to harbour metastatic disease at diagnosis or experience early events, it could complement staging and metabolic burden measures without additional cost or radiation exposure. Materials and methods Study design and cohort This retrospective patient level dataset included 180 patients with breast carcinoma who underwent baseline 18 F-FDG PET/CT prior to definitive therapy. Clinical stage (I to IV), histologic grade (I to III), receptor derived subtype (HR+/HER2+, HR+/HER2-, HR-/HER2+, TNBC), Ki-67 category, and treatment variables (neoadjuvant chemotherapy [NACT], surgery, radiotherapy [RT]) were recorded. ER, PR, HER2 and Ki-67 status were extracted from pathology reports and recorded as binary categories; receptor derived subtype was constructed from hormone receptor status (ER or PR) and HER2 status. PET/CT acquisition and reconstruction parameters All studies were acquired on a GE Discovery IQ 610 (32 slice) hybrid PET/CT system (GE Healthcare, Waukesha, WI, USA). Patients received ~ 3.7 MBq/kg of 18 F-FDG, and imaging was performed after an uptake time of 60 ± 10 minutes. Whole body PET emission data were acquired in 3D mode using a multi bed position protocol (2 minutes per bed position) with bed overlap as per institutional protocol. PET images used for analysis were reconstructed using Q.Clear (BSREM; DICOM reconstruction method QCHD) with CT based attenuation correction, on a 192 × 192 matrix with voxel size 3.65 × 3.65 × 3.23 mm (reconstruction diameter 700 mm). Low dose CT for attenuation correction and anatomical localization was performed at 120 kVp with automatic tube current modulation, reconstructed at 512 × 512 with 0.85 × 0.85 mm in plane pixel spacing and 1.25 mm slice thickness (pitch 0.9375, rotation time 0.7 s; reconstruction diameter 434 mm). Multiplanar reformats (axial, coronal, sagittal) were generated on the workstation. Image analysis and tumour segmentation Primary breast tumour volumes of interest were delineated on baseline PET/CT using a standardised semi-automated workflow with manual refinement and consensus review by two experienced nuclear medicine physicians. To minimise spill-in from physiologic uptake and nodal disease, automated organ and axillary nodal masks generated using TotalSegmentator and an nnU-Net model were used as exclusion priors during contour refinement. Surface mesh extraction and tumor asphericity calculation Tumour sphericity and asphericity were derived from surface area and volume of the segmented primary tumour. Masks were resampled to an isotropic grid for shape analysis and a triangular surface mesh was extracted using Marching cubes. Volume (V) and surface area (A) were computed from the mesh and sphericity (Psi) was calculated as Psi = pi^(1/3) * (6V)^(2/3) / A; asphericity was defined as ASP = 1 - Psi. For modelling, ASP was analysed as a continuous variable (per 0.1 increase) and for Kaplan-Meier analyses patients were dichotomised at the prespecified cut off (0.30). Quantitative metabolic PET metrics SUV based metrics were computed from the native resolution PET image using scanner DICOM metadata for injected dose and patient weight (body weight normalised SUV). Within the primary tumour VOI, SUVmax, metabolic tumour volume (MTV, mL) and total lesion glycolysis (TLG, g) were derived using standard definitions. Endpoints Treatment response was available as a binary variable indicating non-response at the first post treatment assessment. Non-response was defined as absence of an objective response at the first post treatment evaluation as captured in the dataset; because detailed response criteria were not uniformly documented, this endpoint was analysed as exploratory. De novo metastatic disease was defined as stage IV at presentation. Progression free survival (PFS) was defined as time from baseline PET/CT to documented progression or death; overall survival (OS) was defined as time from baseline PET/CT to death from any cause. Patients without an event were administratively censored at last follow up (maximum 84 months). Baseline metastasis was modelled using multivariable logistic regression. Predictive performance was assessed using stratified five-fold cross-validated ROC analysis with out of fold AUC reporting. Model calibration was evaluated using calibration in the large and calibration slope, and by plotting observed versus predicted risk across deciles. To quantify potential optimism in discrimination metrics, bootstrap resampling (300 resamples) was used to derive optimism corrected AUC estimates. PFS and OS were analysed using Kaplan-Meier methods with log-rank testing for ASP groups (cut off 0.30), and using Cox proportional hazards regression. The ASP dichotomisation threshold of 0.30 was selected as a rounded, clinically interpretable value anchored to the cohort specific ROC Youden cut point for baseline metastasis (approximately 0.287) and was then applied consistently across treatment response and survival analyses rather than re optimising a separate threshold for each endpoint. Primary multivariable Cox models included baseline covariates only; recorded treatment variables were evaluated in exploratory analyses and interpreted cautiously because treatments may occur after baseline PET/CT. Because SUVmax, MTV and TLG are correlated measures of metabolic tumour activity, multivariable models used ln(TLG) as the representative metabolic burden variable; ln(SUVmax) and ln(MTV) were evaluated in separate sensitivity analyses and for univariable discrimination. Sensitivity prognostic analyses were additionally performed in patients with stage I to III disease only, with stage modelled as a categorical variable. To assess incremental value of ASP, nested model likelihood ratio tests compared models with and without ASP. All tests were two sided with p < 0.05 considered statistically significant. Clinical utility of baseline metastasis prediction models was assessed using decision curve analysis, estimating net benefit across threshold probabilities from 0.05 to 0.60 and comparing model based strategies with treat all and treat none approaches[20]. Natural logarithm transformation was applied to SUVmax, MTV and TLG for regression modelling; ln denotes natural logarithm. Results Cohort characteristics The analysis included 180 patients (mean age 51.6 ± 11.4 years). Stage distribution was stage I (n = 8 ~ 4.4%), II (n = 74 ~ 41.1%), III (n = 61 ~ 33.9%), and stage IV (n = 37 ~ 20.6%). Receptor derived subtypes were HR+/HER2- (n = 99), TNBC (n = 41), HR-/HER2+ (n = 24), and HR+/HER2+ (n = 16). De novo metastatic disease was present in 37 patients (20.6%). In multivariable logistic regression including age, grade, Ki-67, subtype and ln(TLG), ASP remained an independent predictor of baseline metastasis. ASP (per 0.1 increase) showed OR 3.60 (95% CI 2.02 to 6.43; p < 0.001). ln(TLG) (natural logarithm transformed TLG) was also independently associated with metastasis (OR 2.47; 95% CI 1.32 to 4.63; p = 0.0047). Adding ASP significantly improved model fit compared with the same covariates without ASP (likelihood ratio p < 0.00001). Out of fold AUCs were 0.80 (ASP only), 0.71 (metabolic only), 0.62 (clinical only) and 0.82 (full model). Calibration was acceptable (calibration intercept 0.00; calibration slope 1.00; Brier score 0.117) and is shown in Supplementary Figures S3 and S4. The apparent AUC of the full model was 0.85 and the bootstrap optimism-corrected AUC was 0.81. ROC analysis of ASP alone for baseline metastasis yielded an AUC of 0.804 and identified an optimal cut point of 0.287 by the Youden method (sensitivity 86.5%, specificity 65.7%; Supplementary Figure S5). For clinical interpretability, a rounded cut off of 0.30 was used for Kaplan-Meier stratification. Survival outcomes Kaplan-Meier survival analyses stratified by baseline tumour asphericity using the prespecified cut off (ASP > 0.30 vs ASP ≤ 0.30) demonstrated significantly inferior outcomes in the high asphericity group in the overall cohort (Fig. 3 ), with median overall survival 10.8 months for ASP > 0.30 (n = 71) versus 28.2 months for ASP ≤ 0.30 (n = 109) (log-rank p = 0.00016) and median progression free survival 9.6 versus 24.7 months (log-rank p = 0.0030), with curves administratively censored at 60 months. On molecular subtype stratified analysis (Fig. 4 ), overall survival remained significantly worse with high asphericity in HR+/HER2- (9.7 vs 27.9 months; p = 0.0025) and TNBC (15.1 vs 47.6 months; p = 0.0348), but not in HER2 + disease (10.7 vs 21.3 months; p = 0.268). For progression free survival, high asphericity was associated with shorter PFS in HR+/HER2− (17.4 vs 32.8 months; p = 0.00069), whereas no significant separation was observed in HER2+ (8.8 vs 4.7 months; p = 0.861) or TNBC (4.1 vs 17.7 months; p = 0.932). Separate overall cohort Kaplan-Meier plots for PFS and OS over full follow up are provided as Supplementary Figures S1 and S2. After adjustment for age, clinical stage, grade, subtype, Ki-67 and ln(TLG) in multivariable Cox models, ASP was not independently associated with PFS or OS. For PFS, stage was the dominant predictor (HR 1.86 per stage increment; p < 0.001), and the incremental contribution of ASP was not significant (likelihood ratio p = 0.743). For OS, stage remained the dominant predictor (HR 1.94 per stage increment; p < 0.001), and adding ASP to the clinical plus metabolic model did not improve fit (likelihood ratio p = 0.909). In sensitivity analyses restricted to stage I to III disease with stage modelled categorically, stage III (vs stage I and II) remained associated with shorter PFS (HR 1.69; p = 0.031) and showed a similar trend for OS (HR 1.59; p = 0.077), whereas ASP was not associated with PFS (HR 1.01 per 0.1; p = 0.923) or OS (HR 1.02 per 0.1; p = 0.898). Comparative performance across imaging, clinicopathologic, and biomarker variables is summarised in Figs. 5 and 6 . For baseline metastasis discrimination, ASP provided the highest single variable AUC (0.80), exceeding ln(SUVmax) (0.72), ln(TLG) (0.71), ln(MTV) (0.67) and receptor or Ki-67 variables (AUC ≤ 0.58). For treatment non response, stage (AUC 0.71) and ln(SUVmax) (AUC 0.70) were the strongest discriminators, with ASP showing moderate discrimination (AUC 0.66). For time to event outcomes, stage had the highest Harrell's C for both PFS and OS (0.62 each); among imaging and biomarker variables, ASP had the highest C index (PFS 0.58; OS 0.59) and was associated with higher odds of baseline metastasis (OR 3.70 per 0.1, 95% CI 2.16 to 6.33) and shorter PFS (HR 1.37 per 0.1, 95% CI 1.15 to 1.62) and OS (HR 1.33 per 0.1, 95% CI 1.12 to 1.58). Tables Table 1 Baseline clinicopathologic, treatment, and PET derived tumour characteristics of the study cohort. Continuous variables are presented as mean ± standard deviation or median (interquartile range), as appropriate; categorical variables are presented as n (%). Hormone receptor status and HER2 status are based on standard immunohistochemistry reporting; Ki-67 is reported as percentage Variable Overall Non-metastatic (stage I–III) Metastatic (stage IV) Age 51.6 ± 11.4 52.2 ± 11.1 49.4 ± 12.0 SUVmax 14.1 ± 6.8 13.2 ± 6.6 17.7 ± 6.7 MTV_ml 23.5 ± 22.8 22.3 ± 23.9 28.5 ± 17.2 TLG_g 185.1 ± 169.0 165.9 ± 154.0 259.0 ± 203.5 ASP 0.3 ± 0.1 0.3 ± 0.1 0.3 ± 0.1 PFS_months 26.4 ± 25.6 30.4 ± 26.6 11.2 ± 12.7 OS_months 34.0 ± 31.7 39.5 ± 32.4 12.7 ± 16.2 Stage: I 8 (4.4%) 8 (5.6%) 0 (0.0%) Stage: II 74 (41.1%) 74 (51.7%) 0 (0.0%) Stage: III 61 (33.9%) 61 (42.7%) 0 (0.0%) Stage: IV 37 (20.6%) 0 (0.0%) 37 (100.0%) Grade: I 26 (14.4%) 19 (13.3%) 7 (18.9%) Grade: II 80 (44.4%) 63 (44.1%) 17 (45.9%) Grade: III 74 (41.1%) 61 (42.7%) 13 (35.1%) Subtype_derived: HR+/HER2+ 16 (8.9%) 10 (7.0%) 6 (16.2%) Subtype_derived: HR+/HER2- 99 (55.0%) 85 (59.4%) 14 (37.8%) Subtype_derived: HR-/HER2+ 24 (13.3%) 17 (11.9%) 7 (18.9%) Subtype_derived: TNBC 41 (22.8%) 31 (21.7%) 10 (27.0%) Ki67_high: 0 85 (47.2%) 65 (45.5%) 20 (54.1%) Ki67_high: 1 95 (52.8%) 78 (54.5%) 17 (45.9%) NACT: 0 49 (27.2%) 48 (33.6%) 1 (2.7%) NACT: 1 131 (72.8%) 95 (66.4%) 36 (97.3%) Surgery_any: 0 26 (14.4%) 0 (0.0%) 26 (70.3%) Surgery_any: 1 154 (85.6%) 143 (100.0%) 11 (29.7%) RT: 0 38 (21.1%) 19 (13.3%) 19 (51.4%) RT: 1 142 (78.9%) 124 (86.7%) 18 (48.6%) Table 2 Univariable and multivariable logistic regression for predictors of baseline distant metastasis at presentation. Odds ratios (ORs) are reported with 95% confidence intervals (CI). PET metabolic variables were natural-log transformed (ln) where specified. The multivariable model used a collinearity-safe metabolic burden representation (ln[TLG]) and included tumour asphericity (ASP) to evaluate incremental value beyond clinicopathologic factors. Predictor OR 95% CI p-value Notes Subtype: HR+/HER2- 0.35 0.09 to 1.40 0.137 Subtype: HR-/HER2+ 1.19 0.24 to 5.94 0.836 Subtype: TNBC 1.09 0.26 to 4.56 0.910 ASP (per 0.1) 3.60 2.02 to 6.43 < 0.001 ln(TLG) 2.47 1.32 to 4.63 0.0047 Age 0.98 0.95 to 1.02 0.403 Grade (per level) 0.87 0.47 to 1.62 0.664 Ki-67 high 0.38 0.14 to 1.02 0.054 Table 3 Multivariable Cox proportional hazards regression for progression-free survival. Hazard ratios (HRs) are reported with 95% confidence intervals (CI). PET metabolic variables were natural-log transformed (ln) where specified. Baseline covariates were used for the primary Cox model; treatment variables were evaluated in exploratory analyses and interpreted cautiously. Predictor HR 95% CI p-value Age 1.01 0.99 to 1.02 0.218 Stage (per increment) 1.86 1.42 to 2.45 < 0.001 Grade (per level) 1.05 0.85 to 1.31 0.624 Ki-67 high 1.24 0.88 to 1.75 0.219 ln(TLG) 1.10 0.93 to 1.29 0.270 ASP (per 0.1) 0.96 0.76 to 1.21 0.743 Subtype: HR+/HER2- 0.81 0.46 to 1.46 0.490 Subtype: HR-/HER2+ 1.01 0.52 to 1.95 0.981 Subtype: TNBC 0.93 0.51 to 1.71 0.825 Table 4 Multivariable Cox proportional hazards regression for overall survival. Hazard ratios (HRs) are reported with 95% confidence intervals (CI). PET metabolic variables were natural-log transformed (ln) where specified. Baseline covariates were used for the primary Cox model; treatment variables were evaluated in exploratory analyses and interpreted cautiously. Predictor HR 95% CI p-value Age 1.01 1.00 to 1.03 0.085 Stage (per increment) 1.94 1.46 to 2.59 < 0.001 Grade (per level) 1.26 0.98 to 1.61 0.075 Ki-67 high 1.13 0.77 to 1.64 0.539 ln(TLG) 1.08 0.89 to 1.30 0.428 ASP (per 0.1) 0.99 0.78 to 1.24 0.909 Subtype: HR+/HER2- 1.47 0.79 to 2.76 0.226 Subtype: HR-/HER2+ 1.23 0.60 to 2.52 0.570 Subtype: TNBC 0.96 0.49 to 1.89 0.913 Figures and legends Discussion In our study, baseline tumour asphericity (ASP) was strongly associated with baseline distant metastasis and with shorter PFS and OS on univariable analyses. Notably, ASP provided the strongest single imaging discriminator for baseline metastasis compared with conventional metabolic parameters (SUVmax, MTV and TLG) and biomarker variables, supporting ASP as a compact descriptor of tumour aggressiveness that is not captured by burden metrics alone. For survival endpoints, clinical stage remained the dominant determinant, and ASP did not retain independent prognostic significance after adjustment, suggesting that ASP largely reflects (or co-segregates with) advanced disease extent at presentation. These observations are consistent with biological frameworks that link tumour heterogeneity to acquisition of invasive and metastatic capabilities and microenvironmental pressures such as hypoxia [ 18 , 19 ] . For dichotomised Kaplan-Meier analyses we used ASP 0.30 (ASP > 0.30 vs ASP ≤ 0.30) as a clinically interpretable threshold. This value was not arbitrarily selected, but was anchored to the cohort specific ROC derived Youden cut point for baseline metastasis (0.287; Supplementary Figure S5) and rounded for practical application. ASP remained modelled as a continuous variable in the primary regression analyses, which reduces information loss and limits threshold driven overfitting. From a clinical utility perspective, decision curve analysis indicated that adding ASP to the metastasis prediction model increases net benefit over clinically relevant threshold probabilities, outperforming treat-all and models based on clinical and metabolic burden alone [ 20 ] (Fig. 7 ). If externally validated, an ASP augmented risk model could support risk adapted selection of patients for intensified metastatic work up (or referral for systemic therapy pathways) in settings where whole body PET/CT is not routinely performed for all presentations, without materially increasing the burden of false positive escalation. Beyond baseline metastasis discrimination, baseline tumour asphericity also carried clinically relevant information regarding treatment response. In our cohort, asphericity was associated with higher odds of treatment non response (OR 1.99 per 0.1 increase; p = 0.00033) and provided moderate discrimination for non response (AUC 0.66), comparable to conventional baseline indices such as SUVmax and stage (Figs. 5 and 6 ). This is consistent with prior breast cancer FDG PET evidence that pretreatment imaging phenotype and early treatment related changes can identify patients at risk of suboptimal response and relapse, and with PET/CT radiomics studies in which pretreatment feature signatures predicted pathological response to neoadjuvant chemotherapy and subsequent outcomes [ 21 , 22 , 23 , 24 , 25 ] . A systematic review and meta-analysis supports the prognostic value of FDG PET based response assessment after neoadjuvant chemotherapy in breast cancer [ 29 ] . For survival, higher baseline asphericity (ASP > 0.30) was associated with substantially shorter median overall survival (10.8 vs 28.2 months) and median progression free survival (9.6 vs 24.7 months), and these separations remained evident within molecular subtypes in Kaplan-Meier analyses (Figs. 3 and 4 ). However, these Kaplan-Meier comparisons were unadjusted, and ASP did not remain independently associated with PFS or OS after adjustment for stage, grade, subtype, Ki-67 and ln(TLG) in the multivariable Cox models. This pattern suggests that ASP captures an adverse baseline phenotype and disease extent that is clinically useful for risk stratification, while the dominant adjusted survival signal in this cohort was carried by stage. Prior breast cancer and broader FDG PET studies have reported prognostic associations of pretreatment asphericity and other heterogeneity features, including independent effects in some cohorts, whereas attenuation after adjustment for stage or tumour burden has also been observed depending on cohort composition, endpoint definition and model specification [ 13 , 14 , 26 , 27 , 28 ] . These findings are concordant with prior breast cancer literature demonstrating prognostic relevance of metabolic tumour burden and PET derived heterogeneity. Meta-analyses support MTV and TLG as prognostic biomarkers in breast cancer, although their incremental value can vary by subtype and stage [ 6 , 9 , 10 , 11 , 12 ] . In ER+/HER2- tumours, Groheux and colleagues showed that baseline FDG PET indices have prognostic value, with MTV remaining significant after multivariable adjustment while textural analysis did not add beyond volume measures [ 15 ] . Our results complement these observations by showing that a simple geometric descriptor of FDG uptake, tumour asphericity, captures spatial irregularity distinct from tumour burden, is strongly linked to metastatic presentation and treatment non response, and provides clear unadjusted survival stratification even when stage dominates multivariable Cox models. From a biological standpoint, irregular tumour geometry may reflect infiltrative growth, heterogeneous proliferation, angiogenesis and hypoxia driven selection of aggressive subclones. Hypoxia inducible signalling has been linked to invasion, metastatic dissemination and treatment resistance [ 18 ] , consistent with hallmark mechanisms of cancer progression [ 19 ] . In line with this paradigm, asphericity has shown independent prognostic associations in several FDG avid tumours, including NSCLC and head and neck cancer [ 13 , 14 ] , and more recently in cervical cancer and in a multicentre head and neck cohort [ 27 , 28 ] , supporting generalizability of shape derived risk signals across tumour types. Taken together, baseline tumour asphericity is an interpretable PET derived biomarker that complements stage and conventional metabolic parameters, with value for baseline metastatic risk, treatment non response, and unadjusted survival stratification. In our cohort, the survival association of ASP was substantially attenuated after multivariable adjustment, indicating that its prognostic signal is closely intertwined with disease stage and extent at presentation. Incorporating asphericity into multiparametric baseline models alongside stage, subtype and treatment may still help identify high risk patients who could benefit from intensified systemic therapy, closer surveillance or clinical trial enrolment. Limitations of the study Limitations of the present study include the retrospective design, heterogeneous treatments, and incomplete or non standardized response assessment in some patients. Additionally, asphericity depends on lesion segmentation and can be influenced by image reconstruction, partial volume effects and thresholding. While we applied harmonized workflows and quality control, prospective multicentre validation with standardized response criteria and external calibration of ASP thresholds is needed before routine clinical adoption. A large proportion of patients received neoadjuvant chemotherapy, and primary treatment approach is closely linked to stage and tumour biology; therefore treatment heterogeneity may confound associations with survival and response. In exploratory sensitivity Cox models that additionally adjusted for baseline NACT status, the stage dominant results were unchanged and ASP remained not independently associated with PFS or OS (data not shown). Conclusion Baseline tumour asphericity derived from 18 F-FDG PET/CT is a simple, interpretable marker of tumour geometry and heterogeneity that outperforms conventional metabolic indices and biomarker variables for baseline metastasis discrimination. Higher baseline asphericity was also associated with a higher likelihood of treatment non response and with substantially shorter median PFS and OS on Kaplan-Meier analysis, supporting its utility for baseline risk stratification and clinical communication. However, ASP did not retain independent significance for PFS or OS after adjustment for stage and other baseline covariates in multivariable Cox models in this cohort. Prospective, multicentre validation is warranted. Declarations Funding: None. Conflicts of interest: The authors declare no competing interests. Ethics approval: This retrospective analysis was approved by the Institutional Ethics Committee of Dr Rajendra Prasad Government Medical College, Kangra, Himachal Pradesh, and the requirement for informed consent was waived. Data availability: De-identified data supporting the findings are available from the corresponding author on reasonable request. References Sung H, Ferlay J, Siegel RL et al (2021) Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 71(3):209–249. 10.3322/caac.21660 PMID:33538338 Waks AG, Winer EP (2019) Breast Cancer Treatment: A Review. JAMA. ;321(3):288–300. 10.1001/jama.2018.19323 . PMID:30667505 Arnaout A, Varela NP, Allarakhia M et al (2020) Baseline staging imaging for distant metastasis in women with stages I, II, and III breast cancer. Curr Oncol. ;27(2):e123-e145. 10.3747/co.27.6147 . PMID:32489262 Groheux D, Hindie E, Delord M et al (2016) 18F-FDG PET/CT for staging and restaging of breast cancer. J Nucl Med 57(Suppl 1):17S–26S. 10.2967/jnumed.115.157859 PMID:26834096 Boellaard R, Delgado-Bolton R, Oyen WJG et al (2015) FDG PET/CT: EANM procedure guidelines for tumour imaging: version 2.0. Eur J Nucl Med Mol Imaging 42(2):328–354. 10.1007/s00259-014-2961-x PMID:25452219 Wen W, Xuan D, Hu Y et al (2019) Prognostic value of maximum standard uptake value, metabolic tumor volume, and total lesion glycolysis of positron emission tomography/computed tomography in patients with breast cancer: A systematic review and meta-analysis. PLoS One. ;14(12):e0225959. 10.1371/journal.pone.0225959 . PMID:31826010 Young H, Baum R, Cremerius U et al (1999) Measurement of clinical and subclinical tumour response using [18F]-fluorodeoxyglucose and positron emission tomography: review and 1999 EORTC recommendations. Eur J Cancer 35(13):1773–1782. 10.1016/S0959-8049(99)00229-4 PMID:10673991 Wahl RL, Jacene H, Kasamon Y, Lodge MA (2009) From RECIST to PERCIST: Evolving considerations for PET response criteria in solid tumors. J Nucl Med 50(Suppl 1):122S–150S. 10.2967/jnumed.108.057307 PMID:19403881 Lambin P, Leijenaar RTH, Deist TM et al (2017) Radiomics: the bridge between medical imaging and personalised medicine. Nat Rev Clin Oncol. ;14(12):749–762. 10.1038/nrclinonc.2017.141 . PMID:28975929 Aerts HJWL, Velazquez ER, Leijenaar RTH et al (2014) Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nat Commun 5:4006. 10.1038/ncomms5006 PMID:24892406 Ha S, Choi H, Paeng JC, Cheon GJ (2019) Radiomics in oncological PET/CT: a methodological overview. Nucl Med Mol Imaging 53(1):14–29. 10.1007/s13139-019-00571-4 PMID:30828395 Zwanenburg A, Vallières M, Abdalah MA et al (2020) The Image Biomarker Standardisation Initiative: standardized quantitative radiomics for high-throughput image-based phenotyping. Radiology 295(2):328–338. 10.1148/radiol.2020191145 PMID:32154773 Apostolova I, Rogasch J, Steffen IG et al (2014) Asphericity of pretherapeutic tumor FDG uptake in primary non-small cell lung cancer: a new prognostic marker. BMC Cancer 14:896. 10.1186/1471-2407-14-896 PMID:25444154 Apostolova I, Steffen IG, Wedel F et al (2014) Asphericity of pretherapeutic tumor FDG uptake provides independent prognostic value in head-and-neck cancer. Eur Radiol 24(9):2077–2087. 10.1007/s00330-014-3269-8 PMID:24965509 Groheux D, Martineau A, Teixeira L et al (2017) 18F-FDG PET/CT for predicting the outcome in ER+/HER2- breast cancer patients: comparison of clinicopathological parameters and PET image-derived indices including tumor texture analysis. Breast Cancer Res 19:3. 10.1186/s13058-016-0793-2 PMID:28057031 Wasserthal J, Breit HC, Meyer MT, Pradella M, Hinck D, Sauter AW et al (2023) TotalSegmentator: Robust segmentation of 104 anatomic structures in CT images. Radiol Artif Intell 5(5):e230024. 10.1148/ryai.230024 PMID:37795137 Isensee F, Jaeger PF, Kohl SAA, Petersen J, Maier-Hein KH (2021) nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods 18(2):203–211. 10.1038/s41592-020-01008-z PMID:33288961 Semenza GL (2012) Molecular mechanisms mediating metastasis of hypoxic breast cancer cells. Trends Mol Med. ;18(9):534–543. 10.1016/j.molmed.2012.08.001 . PMID:22921864 Hanahan D, Weinberg RA (2011) Hallmarks of cancer: the next generation. Cell. ;144(5):646–674. 10.1016/j.cell.2011.02.013 . PMID:21376230 Vickers AJ, Elkin EB (2006) Decision curve analysis: a novel method for evaluating prediction models. Med Decis Mak 26(6):565–574. 10.1177/0272989X06295361 PMID:17099194 Groheux D, Sanna A, Majdoub M et al (2015) Baseline tumor 18F-FDG uptake and modifications after 2 cycles of neoadjuvant chemotherapy are prognostic of outcome in ER+/HER2- breast cancer. J Nucl Med 56(6):824–831. 10.2967/jnumed.115.154138 PMID:25883123 Antunovic L, De Sanctis R, Cozzi L et al (2019) PET/CT radiomics in breast cancer: promising tool for prediction of pathological response to neoadjuvant chemotherapy. Eur J Nucl Med Mol Imaging 46(7):1468–1477. 10.1007/s00259-019-04313-8 PMID:30915523 Ha S, Park S, Bang JI, Kim EK, Lee HY (2017) Metabolic radiomics for pretreatment 18F-FDG PET/CT to characterize locally advanced breast cancer: histopathologic characteristics, response to neoadjuvant chemotherapy, and prognosis. Sci Rep 7(1):1556. 10.1038/s41598-017-01524-7 PMID:28484211 Roy S, Whitehead TD, Li S, Ademuyiwa FO, Wahl RL, Dehdashti F, Shoghi KI (2022) Co-clinical FDG-PET radiomic signature in predicting response to neoadjuvant chemotherapy in triple-negative breast cancer. Eur J Nucl Med Mol Imaging. ;49(2):550–562. 10.1007/s00259-021-05489-8 . PMID:34328530. Erratum in: Eur J Nucl Med Mol Imaging. 2022;49(2):788. doi:10.1007/s00259-021-05542-6. PMID:34550428 Chen K, Wang J, Li S et al (2023) Clinical utility of a multi-scanner radiomics model based on 18F-FDG PET/CT for predicting neoadjuvant chemotherapy efficacy in breast cancer: development and external validation. Eur J Nucl Med Mol Imaging 50(7):1869–1880. 10.1007/s00259-023-06150-2 PMID:36808002 Jung JH, Son SH, Kim DH et al (2017) Independent prognostic value of asphericity of pretherapeutic F-18 FDG uptake by primary tumors in patients with breast cancer. Med (Baltim) 96(46):e8438. 10.1097/MD.0000000000008438 PMID:29145250 Cegla P, Hofheinz F, van den Hoff J et al (2023) Asphericity derived from [18F]FDG PET as a new prognostic parameter in cervical cancer patients. Sci Rep 13:8423. 10.1038/s41598-023-35191-8 PMID:37225735 Hausmann P, Sura R, Zschaeck S et al (2025) Tumor asphericity in FDG PET is an independent prognostic parameter improving risk stratification in patients with head and neck squamous cell carcinoma. J Nucl Med 66(5):686–691. 10.2967/jnumed.124.268972 PMID:40081960 Han S, Choi JY (2020) Prognostic value of 18F-FDG PET and PET/CT for assessment of treatment response to neoadjuvant chemotherapy in breast cancer: a systematic review and meta-analysis. Breast Cancer Res 22(1):119. 10.1186/s13058-020-01350-2 PMID:33129348 Additional Declarations The authors declare no competing interests. Supplementary Files SupplementaryFigure.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9221004","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":611890329,"identity":"0347dacc-0429-42a6-955f-218914bcbe3d","order_by":0,"name":"Nitin Gupta","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0klEQVRIiWNgGAWjYDACZiBOYGCQYWBvALIMLIjRwszYANTCw8BzAKRFgihrGEHG8zBIJIB4RGgxZ+c//uDhjjs85pLPr274USDBwN/enYBXi2Uz0JbEM894LGfnlN3sATpM4szZDXi1GBwGaWk7zGNwOyftBg9Qi4FELrFabp5Ju/mHNC032I/dJtYWwxmJbUC/9OSw3ZYxkOAh7JfzBx98/Nl2R86c/fizm2/+2Mjxt/fi1wIFBxgMGHgMQCweYpTDtLA/IFb1KBgFo2AUjDAAAIs0SHOMrH9dAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-9443-9609","institution":"Dr RPGMC Tanda Kangra","correspondingAuthor":true,"prefix":"","firstName":"Nitin","middleName":"","lastName":"Gupta","suffix":""}],"badges":[],"createdAt":"2026-03-25 09:25:35","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9221004/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9221004/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105906031,"identity":"055806ac-84e3-4a1a-8fa2-8a92a7d7d8ce","added_by":"auto","created_at":"2026-04-01 10:16:35","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":477208,"visible":true,"origin":"","legend":"\u003cp\u003eFig1. Receiver operating characteristic (ROC) curves for prediction of baseline metastatic disease and 24 month outcomes using baseline FDG tumour asphericity (ASP) and multivariable models. (A) Baseline distant metastasis at presentation: AUC 0.80 for ASP alone, 0.71 for metabolic model (ln[TLG]), 0.62 for clinicopathologic model, and 0.82 for full model including ASP. (B) 24 month progression free survival (PFS) event: AUC 0.67 for clinicopathologic model, 0.66 for clinicopathologic plus metabolic model, and 0.64 for full model including ASP. (C) 24 month overall survival (OS) event: AUC 0.64 for clinicopathologic model, 0.64 for clinicopathologic plus metabolic model, and 0.63 for full model including ASP.\u003c/p\u003e","description":"","filename":"Figure1ROCpanel.png","url":"https://assets-eu.researchsquare.com/files/rs-9221004/v1/25c637935655684d531332b8.png"},{"id":105873421,"identity":"78893dde-b93f-45b3-91eb-4d7cffb034f9","added_by":"auto","created_at":"2026-04-01 05:12:04","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":287004,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 2\u003c/strong\u003e. Distribution of baseline tumour asphericity (ASP) by clinical stage. ASP increased stepwise with advancing stage (median [IQR] stage I: 0.143 [0.099-0.192], stage II: 0.221 [0.183-0.266], stage III: 0.314 [0.270-0.356], stage IV: 0.346 [0.316-0.377]; Kruskal-Wallis p\u0026lt;0.001).\u003c/p\u003e","description":"","filename":"Figure2ASPbystage.png","url":"https://assets-eu.researchsquare.com/files/rs-9221004/v1/fce0f8ac4ad64114f4a7e47c.png"},{"id":106093213,"identity":"f6d44c9a-232a-48ac-9075-87b06eb9a311","added_by":"auto","created_at":"2026-04-03 11:36:02","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":300118,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 3\u003c/strong\u003e. Kaplan-Meier survival curves stratified by tumour asphericity (ASP \u0026gt;0.30 vs ASP ≤0.30) for the overall cohort, with curves administratively censored at 60 months. Left: overall survival. Right: progression free survival. High asphericity is shown in yellow and low asphericity in red; log-rank p values are shown in each panel. Median OS was 28.2 months (ASP ≤0.30; n=109) versus 10.8 months (ASP \u0026gt;0.30; n=71) (log-rank p = 0.00016). Median PFS was 24.7 months (ASP ≤0.30) versus 9.6 months (ASP \u0026gt;0.30) (log-rank p = 0.0030).\u003c/p\u003e","description":"","filename":"Figure3KMoverallOSPFS60mASP030.png","url":"https://assets-eu.researchsquare.com/files/rs-9221004/v1/ad2f01e7c9e2fd1b6b49b0b1.png"},{"id":105905255,"identity":"fca90d2a-0475-4090-b90c-8521bd5ef23a","added_by":"auto","created_at":"2026-04-01 10:11:47","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":614312,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 4\u003c/strong\u003e. Kaplan-Meier survival curves stratified by tumour asphericity (ASP \u0026gt;0.30 vs ASP ≤0.30) within molecular subtypes, with curves administratively censored at 60 months. Top row: HR+/HER2- (n=99); middle row: HER2+ (n=40); bottom row: triple negative breast cancer (n=41). Left column: overall survival; right column: progression free survival. Log-rank p values are shown in each panel; high asphericity is shown in yellow and low asphericity in red. Median OS (ASP \u0026gt;0.30 vs ASP ≤0.30) was 9.7 vs 27.9 months in HR+/HER2- (p = 0.0025), 10.7 vs 21.3 months in HER2+ (p = 0.268), and 15.1 vs 47.6 months in TNBC (p = 0.0348). Median PFS (ASP \u0026gt;0.30 vs ASP ≤0.30) was 17.4 vs 32.8 months in HR+/HER2- (p = 0.00069), 8.8 vs 4.7 months in HER2+ (p = 0.861), and 4.1 vs 17.7 months in TNBC (p = 0.932).in HR+/HER2- (p = 0.00069), 8.8 vs 4.7 months in HER2+ (p = 0.861), and 4.1 vs 17.7 months in TNBC (p = 0.932).\u003c/p\u003e","description":"","filename":"Figure4KMsubtypesOSPFS60mASP030.png","url":"https://assets-eu.researchsquare.com/files/rs-9221004/v1/e5b3c57225192ee48227294b.png"},{"id":105905385,"identity":"bcdb245e-0756-4850-949e-9ddaf0d28ad0","added_by":"auto","created_at":"2026-04-01 10:11:58","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":619966,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 5\u003c/strong\u003e. Comparative model discrimination of baseline FDG PET asphericity versus metabolic and biomarker variables for baseline metastasis, treatment non response, and survival. Baseline metastasis AUCs: ASP (per 0.1) 0.80, ln(SUVmax) 0.72, ln(TLG) 0.71, ln(MTV) 0.67, receptor and Ki-67 features ≤0.58. Treatment non response AUCs: stage 0.71, ln(SUVmax) 0.70, ASP (per 0.1) 0.66, ln(TLG) 0.62, ln(MTV) 0.58, others ≤0.54. Harrell's C for PFS and OS shows stage as the strongest predictor (0.62 for both) and ASP as the highest performing imaging biomarker (PFS 0.58; OS 0.59). ln denotes natural logarithm transformation.\u003c/p\u003e","description":"","filename":"Figure5Modeldiscriminationpanel.png","url":"https://assets-eu.researchsquare.com/files/rs-9221004/v1/11ac3bdcd96b56ddcc511837.png"},{"id":105905584,"identity":"a2fdc61e-317d-4895-a948-de98015c934b","added_by":"auto","created_at":"2026-04-01 10:12:52","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":679035,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 6\u003c/strong\u003e. Univariable associations of FDG PET tumour asphericity (ASP), PET metabolic parameters, and receptor and Ki-67 status with baseline metastasis, treatment non response, and survival outcomes. Forest plots show odds ratios (OR) for baseline metastasis and treatment non response, and hazard ratios (HR) for PFS and OS, with 95% confidence intervals.\u003c/p\u003e","description":"","filename":"Figure6Effectsizesforestpanel.png","url":"https://assets-eu.researchsquare.com/files/rs-9221004/v1/1b3a26448dbcdb821cb6fbb4.png"},{"id":105873425,"identity":"b78bdd4c-f94c-46c7-acc1-93a89ef86148","added_by":"auto","created_at":"2026-04-01 05:12:04","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":325350,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFigure 7\u003c/strong\u003e. Decision curve analysis (DCA) for prediction of de novo metastatic disease at presentation. Net benefit is shown across threshold probabilities from 0.05 to 0.60 for the clinical model, metabolic model (ln[TLG]), and the ASP augmented model, alongside treat all and treat none strategies; higher curves indicate greater clinical utility.\u003c/p\u003e","description":"","filename":"Figure7DCAmetastasis.png","url":"https://assets-eu.researchsquare.com/files/rs-9221004/v1/787cb1fcd6025b17ff1636d0.png"},{"id":106723869,"identity":"31b33ad0-e116-42cf-bfd5-f201bea3237f","added_by":"auto","created_at":"2026-04-12 18:17:31","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3748924,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9221004/v1/61c95cfa-204c-4278-adda-5c5b4e70e4b9.pdf"},{"id":105873419,"identity":"56a7fdde-c1b0-42e0-852c-27ed284b365d","added_by":"auto","created_at":"2026-04-01 05:12:04","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":775297,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure.docx","url":"https://assets-eu.researchsquare.com/files/rs-9221004/v1/377ae83ee1162f61b6c71365.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eIncremental prognostic value of baseline \u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e18\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003eF-FDG PET/CT tumour asphericity in breast carcinoma: association with de novo metastatic disease and survival outcomes\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eBreast cancer remains the most commonly diagnosed malignancy worldwide and a leading cause of cancer related mortality\u003csup\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]\u003c/sup\u003e. Although major therapeutic advances have improved outcomes, distant metastasis and treatment resistance continue to drive breast cancer deaths \u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/sup\u003e. Baseline staging is therefore essential to guide treatment selection and prognostication, with whole body \u003csup\u003e18\u003c/sup\u003eF‑FDG PET/CT increasingly used in clinically high risk patients to detect extra axillary nodal and distant metastases and to establish a quantitative baseline for response assessment\u003csup\u003e[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/sup\u003e. Beyond detection of overt metastatic disease, baseline PET/CT metrics may help contextualize tumour biology across molecular subtypes and inform risk-adapted systemic and locoregional strategies \u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/sup\u003e .\u003c/p\u003e \u003cp\u003eRobust quantification depends on harmonized acquisition, reconstruction and reporting as emphasized by contemporary procedure guidelines\u003csup\u003e[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e. Conventional PET biomarkers such as SUVmax, metabolic tumour volume (MTV) and total lesion glycolysis (TLG) reflect uptake intensity and tumour burden; meta-analytic evidence supports their prognostic association in breast cancer, but clinically meaningful overlap persists among patients with similar clinicopathologic features\u003csup\u003e[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]\u003c/sup\u003e. Treatment response frameworks including the EORTC recommendations and PERCIST summarize response largely through uptake based measures \u003csup\u003e[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/sup\u003e. Radiomics expands quantitative imaging by extracting descriptors of intensity, texture and geometry that may capture tumour phenotype and microenvironment \u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/sup\u003e. and standardization initiatives such as the Image Biomarker Standardisation Initiative (IBSI) provide consensus definitions to improve reproducibility\u003csup\u003e[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/sup\u003e. Among radiomic feature families, shape metrics are intuitive and comparatively less dependent on intensity normalization; asphericity (ASP\u0026thinsp;=\u0026thinsp;1 - sphericity) summarizes deviation from a sphere and has shown independent prognostic value in other FDG avid cancers beyond conventional metabolic indices\u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]\u003c/sup\u003e. In breast cancer, PET derived radiomic indices including texture features have been linked to outcome, but the incremental value of primary tumour asphericity in routine clinical cohorts remains incompletely defined\u003csup\u003e[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/sup\u003e. Because ASP is derived from surface area and volume, it is readily computable from standard tumour VOIs and is potentially amenable to routine reporting when robust segmentation is available.\u003c/p\u003e \u003cp\u003eThis study aimed to evaluate the prognostic and predictive utility of baseline \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT tumour asphericity (ASP) for baseline distant metastasis, progression free survival (PFS) and overall survival (OS) in breast carcinoma, and to determine whether ASP provides incremental value beyond clinical stage, molecular receptor status (ER/PR/HER2), Ki‑67 and standard metabolic tumour burden metrics (SUVmax, MTV and TLG). We also compared the discriminative performance of ASP against metabolic and biomarker variables for metastasis prediction and assessed its contribution in multivariable survival models.\u003c/p\u003e\n\u003ch3\u003eWhy this matters clinically\u003c/h3\u003e\n\u003cp\u003eEarly risk stratification is central to treatment selection and counselling in breast carcinoma \u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/sup\u003e. Patients presenting with occult metastatic disease often require systemic therapy rather than curative local approaches, and those at high risk of early progression may benefit from intensified therapy and closer surveillance. Baseline distant staging investigations are generally reserved for higher risk presentations\u003csup\u003e[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/sup\u003e and \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT can upstage disease and inform management in intermediate and high risk cohort\u003csup\u003e[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/sup\u003e. If a simple PET derived morphologic metric such as ASP can help identify patients more likely to harbour metastatic disease at diagnosis or experience early events, it could complement staging and metabolic burden measures without additional cost or radiation exposure.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eStudy design and cohort\u003c/h2\u003e \u003cp\u003eThis retrospective patient level dataset included 180 patients with breast carcinoma who underwent baseline \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT prior to definitive therapy. Clinical stage (I to IV), histologic grade (I to III), receptor derived subtype (HR+/HER2+, HR+/HER2-, HR-/HER2+, TNBC), Ki-67 category, and treatment variables (neoadjuvant chemotherapy [NACT], surgery, radiotherapy [RT]) were recorded. ER, PR, HER2 and Ki-67 status were extracted from pathology reports and recorded as binary categories; receptor derived subtype was constructed from hormone receptor status (ER or PR) and HER2 status.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003ePET/CT acquisition and reconstruction parameters\u003c/h3\u003e\n\u003cp\u003eAll studies were acquired on a GE Discovery IQ 610 (32 slice) hybrid PET/CT system (GE Healthcare, Waukesha, WI, USA). Patients received\u0026thinsp;~\u0026thinsp;3.7 MBq/kg of \u003csup\u003e18\u003c/sup\u003eF-FDG, and imaging was performed after an uptake time of 60\u0026thinsp;\u0026plusmn;\u0026thinsp;10 minutes. Whole body PET emission data were acquired in 3D mode using a multi bed position protocol (2 minutes per bed position) with bed overlap as per institutional protocol. PET images used for analysis were reconstructed using Q.Clear (BSREM; DICOM reconstruction method QCHD) with CT based attenuation correction, on a 192 \u0026times; 192 matrix with voxel size 3.65 \u0026times; 3.65 \u0026times; 3.23 mm (reconstruction diameter 700 mm). Low dose CT for attenuation correction and anatomical localization was performed at 120 kVp with automatic tube current modulation, reconstructed at 512 \u0026times; 512 with 0.85 \u0026times; 0.85 mm in plane pixel spacing and 1.25 mm slice thickness (pitch 0.9375, rotation time 0.7 s; reconstruction diameter 434 mm). Multiplanar reformats (axial, coronal, sagittal) were generated on the workstation.\u003c/p\u003e\n\u003ch3\u003eImage analysis and tumour segmentation\u003c/h3\u003e\n\u003cp\u003ePrimary breast tumour volumes of interest were delineated on baseline PET/CT using a standardised semi-automated workflow with manual refinement and consensus review by two experienced nuclear medicine physicians. To minimise spill-in from physiologic uptake and nodal disease, automated organ and axillary nodal masks generated using TotalSegmentator and an nnU-Net model were used as exclusion priors during contour refinement.\u003c/p\u003e\n\u003ch3\u003eSurface mesh extraction and tumor asphericity calculation\u003c/h3\u003e\n\u003cp\u003eTumour sphericity and asphericity were derived from surface area and volume of the segmented primary tumour. Masks were resampled to an isotropic grid for shape analysis and a triangular surface mesh was extracted using Marching cubes. Volume (V) and surface area (A) were computed from the mesh and sphericity (Psi) was calculated as Psi\u0026thinsp;=\u0026thinsp;pi^(1/3) * (6V)^(2/3) / A; asphericity was defined as ASP\u0026thinsp;=\u0026thinsp;1 - Psi. For modelling, ASP was analysed as a continuous variable (per 0.1 increase) and for Kaplan-Meier analyses patients were dichotomised at the prespecified cut off (0.30).\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eQuantitative metabolic PET metrics\u003c/h2\u003e \u003cp\u003eSUV based metrics were computed from the native resolution PET image using scanner DICOM metadata for injected dose and patient weight (body weight normalised SUV). Within the primary tumour VOI, SUVmax, metabolic tumour volume (MTV, mL) and total lesion glycolysis (TLG, g) were derived using standard definitions.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEndpoints\u003c/h3\u003e\n\u003cp\u003eTreatment response was available as a binary variable indicating non-response at the first post treatment assessment. Non-response was defined as absence of an objective response at the first post treatment evaluation as captured in the dataset; because detailed response criteria were not uniformly documented, this endpoint was analysed as exploratory.\u003c/p\u003e \u003cp\u003eDe novo metastatic disease was defined as stage IV at presentation. Progression free survival (PFS) was defined as time from baseline PET/CT to documented progression or death; overall survival (OS) was defined as time from baseline PET/CT to death from any cause. Patients without an event were administratively censored at last follow up (maximum 84 months).\u003c/p\u003e \u003cp\u003eBaseline metastasis was modelled using multivariable logistic regression. Predictive performance was assessed using stratified five-fold cross-validated ROC analysis with out of fold AUC reporting. Model calibration was evaluated using calibration in the large and calibration slope, and by plotting observed versus predicted risk across deciles. To quantify potential optimism in discrimination metrics, bootstrap resampling (300 resamples) was used to derive optimism corrected AUC estimates. PFS and OS were analysed using Kaplan-Meier methods with log-rank testing for ASP groups (cut off 0.30), and using Cox proportional hazards regression. The ASP dichotomisation threshold of 0.30 was selected as a rounded, clinically interpretable value anchored to the cohort specific ROC Youden cut point for baseline metastasis (approximately 0.287) and was then applied consistently across treatment response and survival analyses rather than re optimising a separate threshold for each endpoint. Primary multivariable Cox models included baseline covariates only; recorded treatment variables were evaluated in exploratory analyses and interpreted cautiously because treatments may occur after baseline PET/CT. Because SUVmax, MTV and TLG are correlated measures of metabolic tumour activity, multivariable models used ln(TLG) as the representative metabolic burden variable; ln(SUVmax) and ln(MTV) were evaluated in separate sensitivity analyses and for univariable discrimination. Sensitivity prognostic analyses were additionally performed in patients with stage I to III disease only, with stage modelled as a categorical variable. To assess incremental value of ASP, nested model likelihood ratio tests compared models with and without ASP. All tests were two sided with p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 considered statistically significant. Clinical utility of baseline metastasis prediction models was assessed using decision curve analysis, estimating net benefit across threshold probabilities from 0.05 to 0.60 and comparing model based strategies with treat all and treat none approaches[20]. Natural logarithm transformation was applied to SUVmax, MTV and TLG for regression modelling; ln denotes natural logarithm.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eCohort characteristics\u003c/h2\u003e \u003cp\u003eThe analysis included 180 patients (mean age 51.6\u0026thinsp;\u0026plusmn;\u0026thinsp;11.4 years). Stage distribution was stage I (n\u0026thinsp;=\u0026thinsp;8\u0026thinsp;~\u0026thinsp;4.4%), II (n\u0026thinsp;=\u0026thinsp;74\u0026thinsp;~\u0026thinsp;41.1%), III (n\u0026thinsp;=\u0026thinsp;61\u0026thinsp;~\u0026thinsp;33.9%), and stage IV (n\u0026thinsp;=\u0026thinsp;37\u0026thinsp;~\u0026thinsp;20.6%). Receptor derived subtypes were HR+/HER2- (n\u0026thinsp;=\u0026thinsp;99), TNBC (n\u0026thinsp;=\u0026thinsp;41), HR-/HER2+ (n\u0026thinsp;=\u0026thinsp;24), and HR+/HER2+ (n\u0026thinsp;=\u0026thinsp;16). De novo metastatic disease was present in 37 patients (20.6%).\u003c/p\u003e \u003cp\u003eIn multivariable logistic regression including age, grade, Ki-67, subtype and ln(TLG), ASP remained an independent predictor of baseline metastasis. ASP (per 0.1 increase) showed OR 3.60 (95% CI 2.02 to 6.43; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). ln(TLG) (natural logarithm transformed TLG) was also independently associated with metastasis (OR 2.47; 95% CI 1.32 to 4.63; p\u0026thinsp;=\u0026thinsp;0.0047). Adding ASP significantly improved model fit compared with the same covariates without ASP (likelihood ratio p\u0026thinsp;\u0026lt;\u0026thinsp;0.00001). Out of fold AUCs were 0.80 (ASP only), 0.71 (metabolic only), 0.62 (clinical only) and 0.82 (full model). Calibration was acceptable (calibration intercept 0.00; calibration slope 1.00; Brier score 0.117) and is shown in Supplementary Figures S3 and S4. The apparent AUC of the full model was 0.85 and the bootstrap optimism-corrected AUC was 0.81. ROC analysis of ASP alone for baseline metastasis yielded an AUC of 0.804 and identified an optimal cut point of 0.287 by the Youden method (sensitivity 86.5%, specificity 65.7%; Supplementary Figure S5). For clinical interpretability, a rounded cut off of 0.30 was used for Kaplan-Meier stratification.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eSurvival outcomes\u003c/h2\u003e \u003cp\u003eKaplan-Meier survival analyses stratified by baseline tumour asphericity using the prespecified cut off (ASP\u0026thinsp;\u0026gt;\u0026thinsp;0.30 vs ASP\u0026thinsp;\u0026le;\u0026thinsp;0.30) demonstrated significantly inferior outcomes in the high asphericity group in the overall cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), with median overall survival 10.8 months for ASP\u0026thinsp;\u0026gt;\u0026thinsp;0.30 (n\u0026thinsp;=\u0026thinsp;71) versus 28.2 months for ASP\u0026thinsp;\u0026le;\u0026thinsp;0.30 (n\u0026thinsp;=\u0026thinsp;109) (log-rank p\u0026thinsp;=\u0026thinsp;0.00016) and median progression free survival 9.6 versus 24.7 months (log-rank p\u0026thinsp;=\u0026thinsp;0.0030), with curves administratively censored at 60 months. On molecular subtype stratified analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), overall survival remained significantly worse with high asphericity in HR+/HER2- (9.7 vs 27.9 months; p\u0026thinsp;=\u0026thinsp;0.0025) and TNBC (15.1 vs 47.6 months; p\u0026thinsp;=\u0026thinsp;0.0348), but not in HER2\u0026thinsp;+\u0026thinsp;disease (10.7 vs 21.3 months; p\u0026thinsp;=\u0026thinsp;0.268). For progression free survival, high asphericity was associated with shorter PFS in HR+/HER2\u0026minus; (17.4 vs 32.8 months; p\u0026thinsp;=\u0026thinsp;0.00069), whereas no significant separation was observed in HER2+ (8.8 vs 4.7 months; p\u0026thinsp;=\u0026thinsp;0.861) or TNBC (4.1 vs 17.7 months; p\u0026thinsp;=\u0026thinsp;0.932). Separate overall cohort Kaplan-Meier plots for PFS and OS over full follow up are provided as Supplementary Figures S1 and S2.\u003c/p\u003e \u003cp\u003eAfter adjustment for age, clinical stage, grade, subtype, Ki-67 and ln(TLG) in multivariable Cox models, ASP was not independently associated with PFS or OS. For PFS, stage was the dominant predictor (HR 1.86 per stage increment; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and the incremental contribution of ASP was not significant (likelihood ratio p\u0026thinsp;=\u0026thinsp;0.743). For OS, stage remained the dominant predictor (HR 1.94 per stage increment; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and adding ASP to the clinical plus metabolic model did not improve fit (likelihood ratio p\u0026thinsp;=\u0026thinsp;0.909). In sensitivity analyses restricted to stage I to III disease with stage modelled categorically, stage III (vs stage I and II) remained associated with shorter PFS (HR 1.69; p\u0026thinsp;=\u0026thinsp;0.031) and showed a similar trend for OS (HR 1.59; p\u0026thinsp;=\u0026thinsp;0.077), whereas ASP was not associated with PFS (HR 1.01 per 0.1; p\u0026thinsp;=\u0026thinsp;0.923) or OS (HR 1.02 per 0.1; p\u0026thinsp;=\u0026thinsp;0.898).\u003c/p\u003e \u003cp\u003eComparative performance across imaging, clinicopathologic, and biomarker variables is summarised in Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e and \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. For baseline metastasis discrimination, ASP provided the highest single variable AUC (0.80), exceeding ln(SUVmax) (0.72), ln(TLG) (0.71), ln(MTV) (0.67) and receptor or Ki-67 variables (AUC\u0026thinsp;\u0026le;\u0026thinsp;0.58). For treatment non response, stage (AUC 0.71) and ln(SUVmax) (AUC 0.70) were the strongest discriminators, with ASP showing moderate discrimination (AUC 0.66). For time to event outcomes, stage had the highest Harrell's C for both PFS and OS (0.62 each); among imaging and biomarker variables, ASP had the highest C index (PFS 0.58; OS 0.59) and was associated with higher odds of baseline metastasis (OR 3.70 per 0.1, 95% CI 2.16 to 6.33) and shorter PFS (HR 1.37 per 0.1, 95% CI 1.15 to 1.62) and OS (HR 1.33 per 0.1, 95% CI 1.12 to 1.58).\u003c/p\u003e \u003cp\u003e \u003cb\u003eTables\u003c/b\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline clinicopathologic, treatment, and PET derived tumour characteristics of the study cohort. Continuous variables are presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation or median (interquartile range), as appropriate; categorical variables are presented as n (%). Hormone receptor status and HER2 status are based on standard immunohistochemistry reporting; Ki-67 is reported as percentage\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNon-metastatic (stage I\u0026ndash;III)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMetastatic (stage IV)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e51.6\u0026thinsp;\u0026plusmn;\u0026thinsp;11.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e52.2\u0026thinsp;\u0026plusmn;\u0026thinsp;11.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e49.4\u0026thinsp;\u0026plusmn;\u0026thinsp;12.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSUVmax\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e14.1\u0026thinsp;\u0026plusmn;\u0026thinsp;6.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13.2\u0026thinsp;\u0026plusmn;\u0026thinsp;6.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17.7\u0026thinsp;\u0026plusmn;\u0026thinsp;6.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMTV_ml\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e23.5\u0026thinsp;\u0026plusmn;\u0026thinsp;22.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22.3\u0026thinsp;\u0026plusmn;\u0026thinsp;23.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e28.5\u0026thinsp;\u0026plusmn;\u0026thinsp;17.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTLG_g\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e185.1\u0026thinsp;\u0026plusmn;\u0026thinsp;169.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e165.9\u0026thinsp;\u0026plusmn;\u0026thinsp;154.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e259.0\u0026thinsp;\u0026plusmn;\u0026thinsp;203.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eASP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.3\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.3\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.3\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePFS_months\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26.4\u0026thinsp;\u0026plusmn;\u0026thinsp;25.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e30.4\u0026thinsp;\u0026plusmn;\u0026thinsp;26.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e11.2\u0026thinsp;\u0026plusmn;\u0026thinsp;12.7\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOS_months\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e34.0\u0026thinsp;\u0026plusmn;\u0026thinsp;31.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e39.5\u0026thinsp;\u0026plusmn;\u0026thinsp;32.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e12.7\u0026thinsp;\u0026plusmn;\u0026thinsp;16.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage: I\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8 (4.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e8 (5.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage: II\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e74 (41.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e74 (51.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage: III\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e61 (33.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e61 (42.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage: IV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e37 (20.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e37 (100.0%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade: I\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26 (14.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19 (13.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7 (18.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade: II\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e80 (44.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e63 (44.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17 (45.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade: III\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e74 (41.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e61 (42.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13 (35.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype_derived: HR+/HER2+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e16 (8.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10 (7.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6 (16.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype_derived: HR+/HER2-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e99 (55.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e85 (59.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14 (37.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype_derived: HR-/HER2+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24 (13.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17 (11.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7 (18.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype_derived: TNBC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e41 (22.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e31 (21.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10 (27.0%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKi67_high: 0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e85 (47.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e65 (45.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20 (54.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKi67_high: 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e95 (52.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e78 (54.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17 (45.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNACT: 0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e49 (27.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e48 (33.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1 (2.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNACT: 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e131 (72.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e95 (66.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e36 (97.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSurgery_any: 0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26 (14.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e26 (70.3%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSurgery_any: 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e154 (85.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e143 (100.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e11 (29.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRT: 0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e38 (21.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19 (13.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e19 (51.4%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRT: 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e142 (78.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e124 (86.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e18 (48.6%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eUnivariable and multivariable logistic regression for predictors of baseline distant metastasis at presentation. Odds ratios (ORs) are reported with 95% confidence intervals (CI). PET metabolic variables were natural-log transformed (ln) where specified. The multivariable model used a collinearity-safe metabolic burden representation (ln[TLG]) and included tumour asphericity (ASP) to evaluate incremental value beyond clinicopathologic factors.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e95% CI\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNotes\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype: HR+/HER2-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.09 to 1.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.137\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype: HR-/HER2+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.24 to 5.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.836\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype: TNBC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.26 to 4.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.910\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eASP (per 0.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.02 to 6.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eln(TLG)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.32 to 4.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0047\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.95 to 1.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.403\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade (per level)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.47 to 1.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.664\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKi-67 high\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.14 to 1.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.054\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMultivariable Cox proportional hazards regression for progression-free survival. Hazard ratios (HRs) are reported with 95% confidence intervals (CI). PET metabolic variables were natural-log transformed (ln) where specified. Baseline covariates were used for the primary Cox model; treatment variables were evaluated in exploratory analyses and interpreted cautiously.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e95% CI\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.99 to 1.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.218\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage (per increment)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.42 to 2.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade (per level)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.85 to 1.31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.624\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKi-67 high\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.88 to 1.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.219\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eln(TLG)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.93 to 1.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.270\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eASP (per 0.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.76 to 1.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.743\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype: HR+/HER2-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.46 to 1.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.490\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype: HR-/HER2+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.52 to 1.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.981\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype: TNBC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.93\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.51 to 1.71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.825\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMultivariable Cox proportional hazards regression for overall survival. Hazard ratios (HRs) are reported with 95% confidence intervals (CI). PET metabolic variables were natural-log transformed (ln) where specified. Baseline covariates were used for the primary Cox model; treatment variables were evaluated in exploratory analyses and interpreted cautiously.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e95% CI\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.00 to 1.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.085\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage (per increment)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.46 to 2.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGrade (per level)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.98 to 1.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.075\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKi-67 high\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.77 to 1.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.539\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eln(TLG)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.89 to 1.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.428\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eASP (per 0.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.78 to 1.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.909\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype: HR+/HER2-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.79 to 2.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.226\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype: HR-/HER2+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.60 to 2.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.570\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubtype: TNBC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.49 to 1.89\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.913\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eFigures and legends\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn our study, baseline tumour asphericity (ASP) was strongly associated with baseline distant metastasis and with shorter PFS and OS on univariable analyses. Notably, ASP provided the strongest single imaging discriminator for baseline metastasis compared with conventional metabolic parameters (SUVmax, MTV and TLG) and biomarker variables, supporting ASP as a compact descriptor of tumour aggressiveness that is not captured by burden metrics alone. For survival endpoints, clinical stage remained the dominant determinant, and ASP did not retain independent prognostic significance after adjustment, suggesting that ASP largely reflects (or co-segregates with) advanced disease extent at presentation. These observations are consistent with biological frameworks that link tumour heterogeneity to acquisition of invasive and metastatic capabilities and microenvironmental pressures such as hypoxia\u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eFor dichotomised Kaplan-Meier analyses we used ASP 0.30 (ASP\u0026thinsp;\u0026gt;\u0026thinsp;0.30 vs ASP\u0026thinsp;\u0026le;\u0026thinsp;0.30) as a clinically interpretable threshold. This value was not arbitrarily selected, but was anchored to the cohort specific ROC derived Youden cut point for baseline metastasis (0.287; Supplementary Figure S5) and rounded for practical application. ASP remained modelled as a continuous variable in the primary regression analyses, which reduces information loss and limits threshold driven overfitting.\u003c/p\u003e \u003cp\u003eFrom a clinical utility perspective, decision curve analysis indicated that adding ASP to the metastasis prediction model increases net benefit over clinically relevant threshold probabilities, outperforming treat-all and models based on clinical and metabolic burden alone\u003csup\u003e[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/sup\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e). If externally validated, an ASP augmented risk model could support risk adapted selection of patients for intensified metastatic work up (or referral for systemic therapy pathways) in settings where whole body PET/CT is not routinely performed for all presentations, without materially increasing the burden of false positive escalation.\u003c/p\u003e \u003cp\u003eBeyond baseline metastasis discrimination, baseline tumour asphericity also carried clinically relevant information regarding treatment response. In our cohort, asphericity was associated with higher odds of treatment non response (OR 1.99 per 0.1 increase; p\u0026thinsp;=\u0026thinsp;0.00033) and provided moderate discrimination for non response (AUC 0.66), comparable to conventional baseline indices such as SUVmax and stage (Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e and \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). This is consistent with prior breast cancer FDG PET evidence that pretreatment imaging phenotype and early treatment related changes can identify patients at risk of suboptimal response and relapse, and with PET/CT radiomics studies in which pretreatment feature signatures predicted pathological response to neoadjuvant chemotherapy and subsequent outcomes\u003csup\u003e[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]\u003c/sup\u003e. A systematic review and meta-analysis supports the prognostic value of FDG PET based response assessment after neoadjuvant chemotherapy in breast cancer\u003csup\u003e[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eFor survival, higher baseline asphericity (ASP\u0026thinsp;\u0026gt;\u0026thinsp;0.30) was associated with substantially shorter median overall survival (10.8 vs 28.2 months) and median progression free survival (9.6 vs 24.7 months), and these separations remained evident within molecular subtypes in Kaplan-Meier analyses (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). However, these Kaplan-Meier comparisons were unadjusted, and ASP did not remain independently associated with PFS or OS after adjustment for stage, grade, subtype, Ki-67 and ln(TLG) in the multivariable Cox models. This pattern suggests that ASP captures an adverse baseline phenotype and disease extent that is clinically useful for risk stratification, while the dominant adjusted survival signal in this cohort was carried by stage. Prior breast cancer and broader FDG PET studies have reported prognostic associations of pretreatment asphericity and other heterogeneity features, including independent effects in some cohorts, whereas attenuation after adjustment for stage or tumour burden has also been observed depending on cohort composition, endpoint definition and model specification\u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThese findings are concordant with prior breast cancer literature demonstrating prognostic relevance of metabolic tumour burden and PET derived heterogeneity. Meta-analyses support MTV and TLG as prognostic biomarkers in breast cancer, although their incremental value can vary by subtype and stage\u003csup\u003e[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/sup\u003e. In ER+/HER2- tumours, Groheux and colleagues showed that baseline FDG PET indices have prognostic value, with MTV remaining significant after multivariable adjustment while textural analysis did not add beyond volume measures\u003csup\u003e[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/sup\u003e. Our results complement these observations by showing that a simple geometric descriptor of FDG uptake, tumour asphericity, captures spatial irregularity distinct from tumour burden, is strongly linked to metastatic presentation and treatment non response, and provides clear unadjusted survival stratification even when stage dominates multivariable Cox models.\u003c/p\u003e \u003cp\u003eFrom a biological standpoint, irregular tumour geometry may reflect infiltrative growth, heterogeneous proliferation, angiogenesis and hypoxia driven selection of aggressive subclones. Hypoxia inducible signalling has been linked to invasion, metastatic dissemination and treatment resistance\u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e, consistent with hallmark mechanisms of cancer progression\u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e. In line with this paradigm, asphericity has shown independent prognostic associations in several FDG avid tumours, including NSCLC and head and neck cancer \u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]\u003c/sup\u003e, and more recently in cervical cancer and in a multicentre head and neck cohort\u003csup\u003e[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]\u003c/sup\u003e, supporting generalizability of shape derived risk signals across tumour types. Taken together, baseline tumour asphericity is an interpretable PET derived biomarker that complements stage and conventional metabolic parameters, with value for baseline metastatic risk, treatment non response, and unadjusted survival stratification. In our cohort, the survival association of ASP was substantially attenuated after multivariable adjustment, indicating that its prognostic signal is closely intertwined with disease stage and extent at presentation. Incorporating asphericity into multiparametric baseline models alongside stage, subtype and treatment may still help identify high risk patients who could benefit from intensified systemic therapy, closer surveillance or clinical trial enrolment.\u003c/p\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eLimitations of the study\u003c/h2\u003e \u003cp\u003eLimitations of the present study include the retrospective design, heterogeneous treatments, and incomplete or non standardized response assessment in some patients. Additionally, asphericity depends on lesion segmentation and can be influenced by image reconstruction, partial volume effects and thresholding. While we applied harmonized workflows and quality control, prospective multicentre validation with standardized response criteria and external calibration of ASP thresholds is needed before routine clinical adoption. A large proportion of patients received neoadjuvant chemotherapy, and primary treatment approach is closely linked to stage and tumour biology; therefore treatment heterogeneity may confound associations with survival and response. In exploratory sensitivity Cox models that additionally adjusted for baseline NACT status, the stage dominant results were unchanged and ASP remained not independently associated with PFS or OS (data not shown).\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eBaseline tumour asphericity derived from \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT is a simple, interpretable marker of tumour geometry and heterogeneity that outperforms conventional metabolic indices and biomarker variables for baseline metastasis discrimination. Higher baseline asphericity was also associated with a higher likelihood of treatment non response and with substantially shorter median PFS and OS on Kaplan-Meier analysis, supporting its utility for baseline risk stratification and clinical communication. However, ASP did not retain independent significance for PFS or OS after adjustment for stage and other baseline covariates in multivariable Cox models in this cohort. Prospective, multicentre validation is warranted.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eFunding: None.\u003c/p\u003e\n\u003cp\u003eConflicts of interest: The authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003eEthics approval: This retrospective analysis was approved by the Institutional Ethics Committee of Dr Rajendra Prasad Government Medical College, Kangra, Himachal Pradesh, and the requirement for informed consent was waived.\u003c/p\u003e\n\u003cp\u003eData availability: De-identified data supporting the findings are available from the corresponding author on reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSung H, Ferlay J, Siegel RL et al (2021) Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 71(3):209\u0026ndash;249. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3322/caac.21660\u003c/span\u003e\u003cspan address=\"10.3322/caac.21660\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:33538338\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWaks AG, Winer EP (2019) Breast Cancer Treatment: A Review. JAMA. ;321(3):288\u0026ndash;300. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jama.2018.19323\u003c/span\u003e\u003cspan address=\"10.1001/jama.2018.19323\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID:30667505\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArnaout A, Varela NP, Allarakhia M et al (2020) Baseline staging imaging for distant metastasis in women with stages I, II, and III breast cancer. Curr Oncol. ;27(2):e123-e145. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3747/co.27.6147\u003c/span\u003e\u003cspan address=\"10.3747/co.27.6147\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID:32489262\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGroheux D, Hindie E, Delord M et al (2016) 18F-FDG PET/CT for staging and restaging of breast cancer. J Nucl Med 57(Suppl 1):17S\u0026ndash;26S. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2967/jnumed.115.157859\u003c/span\u003e\u003cspan address=\"10.2967/jnumed.115.157859\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:26834096\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoellaard R, Delgado-Bolton R, Oyen WJG et al (2015) FDG PET/CT: EANM procedure guidelines for tumour imaging: version 2.0. Eur J Nucl Med Mol Imaging 42(2):328\u0026ndash;354. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00259-014-2961-x\u003c/span\u003e\u003cspan address=\"10.1007/s00259-014-2961-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:25452219\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWen W, Xuan D, Hu Y et al (2019) Prognostic value of maximum standard uptake value, metabolic tumor volume, and total lesion glycolysis of positron emission tomography/computed tomography in patients with breast cancer: A systematic review and meta-analysis. PLoS One. ;14(12):e0225959. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0225959\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0225959\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID:31826010\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoung H, Baum R, Cremerius U et al (1999) Measurement of clinical and subclinical tumour response using [18F]-fluorodeoxyglucose and positron emission tomography: review and 1999 EORTC recommendations. Eur J Cancer 35(13):1773\u0026ndash;1782. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S0959-8049(99)00229-4\u003c/span\u003e\u003cspan address=\"10.1016/S0959-8049(99)00229-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:10673991\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWahl RL, Jacene H, Kasamon Y, Lodge MA (2009) From RECIST to PERCIST: Evolving considerations for PET response criteria in solid tumors. J Nucl Med 50(Suppl 1):122S\u0026ndash;150S. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2967/jnumed.108.057307\u003c/span\u003e\u003cspan address=\"10.2967/jnumed.108.057307\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:19403881\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLambin P, Leijenaar RTH, Deist TM et al (2017) Radiomics: the bridge between medical imaging and personalised medicine. Nat Rev Clin Oncol. ;14(12):749\u0026ndash;762. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nrclinonc.2017.141\u003c/span\u003e\u003cspan address=\"10.1038/nrclinonc.2017.141\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID:28975929\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAerts HJWL, Velazquez ER, Leijenaar RTH et al (2014) Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nat Commun 5:4006. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/ncomms5006\u003c/span\u003e\u003cspan address=\"10.1038/ncomms5006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:24892406\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHa S, Choi H, Paeng JC, Cheon GJ (2019) Radiomics in oncological PET/CT: a methodological overview. Nucl Med Mol Imaging 53(1):14\u0026ndash;29. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s13139-019-00571-4\u003c/span\u003e\u003cspan address=\"10.1007/s13139-019-00571-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:30828395\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZwanenburg A, Valli\u0026egrave;res M, Abdalah MA et al (2020) The Image Biomarker Standardisation Initiative: standardized quantitative radiomics for high-throughput image-based phenotyping. Radiology 295(2):328\u0026ndash;338. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1148/radiol.2020191145\u003c/span\u003e\u003cspan address=\"10.1148/radiol.2020191145\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:32154773\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eApostolova I, Rogasch J, Steffen IG et al (2014) Asphericity of pretherapeutic tumor FDG uptake in primary non-small cell lung cancer: a new prognostic marker. BMC Cancer 14:896. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/1471-2407-14-896\u003c/span\u003e\u003cspan address=\"10.1186/1471-2407-14-896\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:25444154\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eApostolova I, Steffen IG, Wedel F et al (2014) Asphericity of pretherapeutic tumor FDG uptake provides independent prognostic value in head-and-neck cancer. Eur Radiol 24(9):2077\u0026ndash;2087. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00330-014-3269-8\u003c/span\u003e\u003cspan address=\"10.1007/s00330-014-3269-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:24965509\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGroheux D, Martineau A, Teixeira L et al (2017) 18F-FDG PET/CT for predicting the outcome in ER+/HER2- breast cancer patients: comparison of clinicopathological parameters and PET image-derived indices including tumor texture analysis. Breast Cancer Res 19:3. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13058-016-0793-2\u003c/span\u003e\u003cspan address=\"10.1186/s13058-016-0793-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:28057031\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWasserthal J, Breit HC, Meyer MT, Pradella M, Hinck D, Sauter AW et al (2023) TotalSegmentator: Robust segmentation of 104 anatomic structures in CT images. Radiol Artif Intell 5(5):e230024. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1148/ryai.230024\u003c/span\u003e\u003cspan address=\"10.1148/ryai.230024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:37795137\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIsensee F, Jaeger PF, Kohl SAA, Petersen J, Maier-Hein KH (2021) nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods 18(2):203\u0026ndash;211. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41592-020-01008-z\u003c/span\u003e\u003cspan address=\"10.1038/s41592-020-01008-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:33288961\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSemenza GL (2012) Molecular mechanisms mediating metastasis of hypoxic breast cancer cells. Trends Mol Med. ;18(9):534\u0026ndash;543. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.molmed.2012.08.001\u003c/span\u003e\u003cspan address=\"10.1016/j.molmed.2012.08.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID:22921864\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanahan D, Weinberg RA (2011) Hallmarks of cancer: the next generation. Cell. ;144(5):646\u0026ndash;674. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cell.2011.02.013\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2011.02.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID:21376230\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVickers AJ, Elkin EB (2006) Decision curve analysis: a novel method for evaluating prediction models. Med Decis Mak 26(6):565\u0026ndash;574. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1177/0272989X06295361\u003c/span\u003e\u003cspan address=\"10.1177/0272989X06295361\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:17099194\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGroheux D, Sanna A, Majdoub M et al (2015) Baseline tumor 18F-FDG uptake and modifications after 2 cycles of neoadjuvant chemotherapy are prognostic of outcome in ER+/HER2- breast cancer. J Nucl Med 56(6):824\u0026ndash;831. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2967/jnumed.115.154138\u003c/span\u003e\u003cspan address=\"10.2967/jnumed.115.154138\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:25883123\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAntunovic L, De Sanctis R, Cozzi L et al (2019) PET/CT radiomics in breast cancer: promising tool for prediction of pathological response to neoadjuvant chemotherapy. Eur J Nucl Med Mol Imaging 46(7):1468\u0026ndash;1477. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00259-019-04313-8\u003c/span\u003e\u003cspan address=\"10.1007/s00259-019-04313-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:30915523\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHa S, Park S, Bang JI, Kim EK, Lee HY (2017) Metabolic radiomics for pretreatment 18F-FDG PET/CT to characterize locally advanced breast cancer: histopathologic characteristics, response to neoadjuvant chemotherapy, and prognosis. Sci Rep 7(1):1556. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-017-01524-7\u003c/span\u003e\u003cspan address=\"10.1038/s41598-017-01524-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:28484211\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoy S, Whitehead TD, Li S, Ademuyiwa FO, Wahl RL, Dehdashti F, Shoghi KI (2022) Co-clinical FDG-PET radiomic signature in predicting response to neoadjuvant chemotherapy in triple-negative breast cancer. Eur J Nucl Med Mol Imaging. ;49(2):550\u0026ndash;562. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00259-021-05489-8\u003c/span\u003e\u003cspan address=\"10.1007/s00259-021-05489-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID:34328530. Erratum in: Eur J Nucl Med Mol Imaging. 2022;49(2):788. doi:10.1007/s00259-021-05542-6. PMID:34550428\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen K, Wang J, Li S et al (2023) Clinical utility of a multi-scanner radiomics model based on 18F-FDG PET/CT for predicting neoadjuvant chemotherapy efficacy in breast cancer: development and external validation. Eur J Nucl Med Mol Imaging 50(7):1869\u0026ndash;1880. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00259-023-06150-2\u003c/span\u003e\u003cspan address=\"10.1007/s00259-023-06150-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:36808002\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJung JH, Son SH, Kim DH et al (2017) Independent prognostic value of asphericity of pretherapeutic F-18 FDG uptake by primary tumors in patients with breast cancer. Med (Baltim) 96(46):e8438. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/MD.0000000000008438\u003c/span\u003e\u003cspan address=\"10.1097/MD.0000000000008438\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:29145250\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCegla P, Hofheinz F, van den Hoff J et al (2023) Asphericity derived from [18F]FDG PET as a new prognostic parameter in cervical cancer patients. Sci Rep 13:8423. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-023-35191-8\u003c/span\u003e\u003cspan address=\"10.1038/s41598-023-35191-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:37225735\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHausmann P, Sura R, Zschaeck S et al (2025) Tumor asphericity in FDG PET is an independent prognostic parameter improving risk stratification in patients with head and neck squamous cell carcinoma. J Nucl Med 66(5):686\u0026ndash;691. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2967/jnumed.124.268972\u003c/span\u003e\u003cspan address=\"10.2967/jnumed.124.268972\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:40081960\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHan S, Choi JY (2020) Prognostic value of 18F-FDG PET and PET/CT for assessment of treatment response to neoadjuvant chemotherapy in breast cancer: a systematic review and meta-analysis. Breast Cancer Res 22(1):119. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13058-020-01350-2\u003c/span\u003e\u003cspan address=\"10.1186/s13058-020-01350-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID:33129348\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"18F-FDG PET/CT, breast cancer, tumour asphericity, metabolic tumour volume, total lesion glycolysis, metastasis, progression free survival, overall survival","lastPublishedDoi":"10.21203/rs.3.rs-9221004/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9221004/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eTumour asphericity (ASP), derived from baseline 18F-FDG PET/CT, may capture morphologic heterogeneity linked to aggressive biology and adverse outcomes.\u003c/p\u003e\u003ch2\u003eAim\u003c/h2\u003e \u003cp\u003eTo evaluate the incremental value of baseline ASP for de novo metastatic disease at presentation and survival outcomes in breast carcinoma beyond clinical stage, receptor-derived subtype, Ki-67 and metabolic tumour burden.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003ePatient-level data (n\u0026thinsp;=\u0026thinsp;180) were analysed. De novo metastasis (stage IV) was modelled using multivariable logistic regression with 5-fold cross-validated ROC analysis; model calibration was assessed using observed versus predicted risk. PFS and OS were analysed using Kaplan-Meier, log-rank tests and multivariable Cox proportional hazards regression.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eDe novo metastatic disease was present in 37/180 patients (20.6%). In multivariable logistic regression, ASP (per 0.1 increase) independently predicted baseline metastasis (OR 3.60, 95% CI 2.02 to 6.43, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) along with ln(TLG) (natural logarithm transformed TLG; OR 2.47, 95% CI 1.32 to 4.63, p\u0026thinsp;=\u0026thinsp;0.0047). Out-of-fold AUC for metastasis prediction was 0.80 for ASP alone and 0.82 for the full model including ASP; calibration was satisfactory (calibration slope 1.00; Brier score 0.117). Higher ASP was also associated with treatment non-response (OR 1.99 per 0.1, p\u0026thinsp;=\u0026thinsp;0.00033). On unadjusted analysis over full follow-up (not administratively censored), ASP\u0026thinsp;\u0026gt;\u0026thinsp;0.30 (vs ASP\u0026thinsp;\u0026le;\u0026thinsp;0.30) was associated with shorter PFS (median 24.7 vs 9.6 months; log-rank p\u0026thinsp;=\u0026thinsp;0.0017) and OS (median 28.2 vs 10.8 months; log-rank p\u0026thinsp;=\u0026thinsp;0.00011).\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eBaseline tumour asphericity derived from \u003csup\u003e18\u003c/sup\u003eF-FDG PET/CT is a simple, interpretable geometric biomarker that outperforms conventional metabolic indices and biomarker variables for baseline metastasis discrimination and is associated with treatment non-response and shorter median PFS and OS on unadjusted survival analysis. Prospective multicentre validation is warranted.\u003c/p\u003e","manuscriptTitle":"Incremental prognostic value of baseline 18F-FDG PET/CT tumour asphericity in breast carcinoma: association with de novo metastatic disease and survival outcomes","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-01 05:12:00","doi":"10.21203/rs.3.rs-9221004/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"7a534198-291d-4697-b072-7bdd8e6666da","owner":[],"postedDate":"April 1st, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":65102221,"name":"Nuclear Medicine \u0026 Medical Imaging"}],"tags":[],"updatedAt":"2026-04-01T05:12:00+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-01 05:12:00","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9221004","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9221004","identity":"rs-9221004","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}