CT and MRI in Advanced Ovarian Cancer: Advances in Imaging Techniques.

Imaging is pivotal in evaluating tumor location, size, and extent of invasion into adjacent structures, which directly influence treatment decisions. Initial imaging typically involves a pelvic ultrasound scan, but approximately 20% of solitary adnexal masses remain sonographically indeterminate. For these patients, mp-MRI is the imaging modality of choice due to its superior image characterization capabilities [ 9 ]. CE-CT remains standard for staging and monitoring of treatment response in patients with suspected or confirmed OC [ 10 ]. Although CE-CT demonstrates high specificity for detecting tumors, its sensitivity in identifying tumors at critical surgical sites, such as peritoneal metastases, LN, and other areas, remains limited [ 11 ]. More recently, mp-MRI and PET-CT have been increasingly used in specialist centers for staging, providing complementary functional assessments to guide operability and treatment decisions. The Ovarian-Adnexal Reporting and Data System for MRI (O-RADS MRI) has been developed to provide standardized risk stratification for adnexal masses that remain indeterminate on ultrasound. This system classifies adnexal masses according to the malignancy risk based on lesion morphology, enhancement patterns, and the presence of solid components. It improves diagnostic consistency and clinical decision-making, with high sensitivity (93%) and specificity (91%) for detecting malignancy [ 12 ]. In cases of suspected advanced OC or peritoneal disease, lesions meeting the criteria for O-RADS 5 are considered highly suspicious for malignancy (positive predictive value [PPV] 90%). This classification supports the identification of patients who require more aggressive management and facilitates surgical planning, particularly when the disease involves complex sites such as the suprarenal LNs or extensive peritoneal surfaces [ 13 ]. Currently, the European Society of Gynaecological Oncology, the European Society for Medical Oncology, and the European Society of Pathology recommend comprehensive staging using either CE-CT or diffusion-weighted imaging (DWI) MRI to assess tumor burden and location [ 7 ]. Areas that are particularly challenging to resect, such as suprarenal LN and extensive peritoneal metastases, must be thoroughly evaluated as they may complicate the surgical approach [ 14 ]. Furthermore, despite advances in imaging techniques, disparities in protocols and equipment persist within and between healthcare institutions, which can impact the quality of OC assessment. Efforts have been made to develop optimal imaging protocols ( Table 1 ), with the sensitivities and specificities summarized in Table 2 . Thus, a multidisciplinary approach integrating a specialist radiologist with high-quality imaging and discussion in specialist gynecologist tumor board is essential in the stratification of patients into the most appropriate management.

Ovarian cancer (OC) remains one of the leading causes of gynecologic cancer-related mortality, accounting for approximately 200000 deaths annually worldwide [ 1 ]. The lack of early symptoms and effective screening methods results in late-stage diagnoses, often at FIGO Stage III–IV, when metastasis to the peritoneum and distant organs has already occurred [ 2 3 ]. The International Federation of Gynaecology and Obstetrics (FIGO) surgical staging remains the gold standard for OC classification [ 4 ]. Epithelial OC (EOC), the most common OC subtype, accounts for approximately 90% of cases [ 4 5 ]. Advanced OC treatment routinely includes a combination of cytoreductive surgery (CRS), platinum-based chemotherapy, and targeted maintenance therapies based on the genetic profile of the patient and tumor. The timing of chemotherapy (i.e., neoadjuvant or adjuvant) depends on disease extent, surgical resectability, performance status, comorbidities, and patient's wishes [ 6 ]. CRS typically involves total abdominal hysterectomy, bilateral salpingo-oophorectomy, and infragastric omentectomy, with additional procedures as necessary to complete tumor clearance, depending on the tumor dissemination patterns. These may include splenectomy, bowel resection, and removal of bulky lymph nodes (LNs), visceral and parietal peritoneum, and other organs [ 7 ]. Cytoreductive procedures may also extend extra-abdominally (e.g., paracardiac, inguinal, and axillary areas and pleural space). Achieving complete macroscopic tumor clearance in the entire body is considered the most crucial prognostic factor for survival, and CRS should only be performed if complete macroscopic resection is thought to be achievable with an acceptable morbidity profile [ 7 8 ]. Accurate preoperative imaging is key for adequate surgical planning and obtaining informed patient consent. This pictorial review highlights the current role of contrast-enhanced computed tomography (CE-CT) and multiparametric magnetic resonance imaging (mp-MRI) in advanced OC and its impact on clinical management.

LN involvement is paramount in the staging and prognosis of OC, as it is often associated with advanced disease and poorer prognosis ( Figs. 14 , 15 ). CE-CT has limited diagnostic accuracy in malignant LN detection with a 1 cm short-axis threshold. Overall, sensitivity was 40.7%, and specificity was 89.1% [ 29 30 ]. Routine systematic pelvic and para-aortic lymphadenectomy is not recommended in patients who have undergone complete tumor resection and have normal LNs pre- and intraoperatively, as it has not demonstrated a survival benefit. Instead, the targeted removal of abnormal LNs is advised when clinically indicated and complete peritoneal clearance has otherwise been achieved [ 29 ]. CE-CT is more precise for detecting para-aortic LN (sensitivity, 41.2%; specificity, 93.1%) than pelvic LN (sensitivity, 26.6%; specificity, 91.8%) [ 30 ]. A meta-analysis has shown that FDG-PET-CT outperforms CE-CT in identifying malignant LN, with a higher sensitivity (73%) and specificity (97%) than CE-CT (43% sensitivity, 95% specificity) and mp-MRI (55% sensitivity, 88% specificity) [ 31 ]. Malignant cardiophrenic LNs are typically assessed on CE-CT using a short-axis threshold >7 mm, yielding an 86% PPV [ 32 ]. However, a study comparing CE-CT and gated FDG-PET-CT to detect thoracic involvement in advanced OC concluded that PET-CT improved diagnostic accuracy, upstaging 46% of the cases and downstaging 8%, but only altered clinical management in 5% of the patients [ 33 ]. The imaging review areas included the supraclavicular, axillary, and internal mammary LN chains [ 34 ].

Metastasis can involve the pleura ( Fig. 12 ), lungs ( Fig. 13 ), axilla, inguinal area, and mediastinum, including the paracardiac space. The detection of such lesions is critical in determining the appropriateness of radical cytoreduction over less radical nonsurgical treatment options. Mironov et al. [ 27 ] analyzed CE-CT features and found that pleural effusion and ascites were associated with solid pleural disease. Other suspicious pleural features include nodules, thickening (>3 mm), and enhancement ( Figs. 12 , 13 ). Another study evaluated patients with EOC preoperatively using WB-DWI MRI, which accurately identified unresectable sites compared with CE-CT, particularly when surgical PCI was >20 [ 28 ].

Recent efforts to standardize and improve OC imaging interpretation have led to the development of the OC reporting Lexicon for CE-CT and mp-MRI [ 34 ]. The lexicon provides a framework for imaging descriptors of OC. It incorporates various imaging features (e.g., general, adnexal lesion, peritoneal carcinomatosis, LN, metastatic disease, and fluid) to unify the terms used by radiologists during reporting [ 40 41 ]. This structured method for evaluating OC and communicating findings among the multidisciplinary team helps guide management decisions, especially in the preoperative setting.

Advancements in imaging technologies and artificial intelligence (AI) hold promise for improving personalized management of advanced OC, although not currently implemented in daily practice. AI techniques, including machine-learning and deep-learning algorithms, have demonstrated strong diagnostic performance in detecting OC across a range of imaging platforms, including ultrasound, CT, and MRI. A recent meta-analysis of 28 studies reported a high pooled sensitivity (88%) and specificity (85%), with an AUC of 0.93. Machine-learning models slightly outperformed deep-learning models in both sensitivity and specificity. Subgroup analyses showed consistent diagnostic performance across different imaging modalities (ultrasound, MRI, and CT), geographic regions (Asia and non-Asia), sample sizes, and comparisons with radiologists [ 42 ]. Radiomics, which extracts quantitative imaging features from medical scans, may enhance the precision of metastasis detection. Studies have demonstrated that integrated CT-based radiomics models can predict resectability (AUC 0.881 vs. 0.675) and metastatic risk (AUC 0.909 vs. 0.860) more accurately than clinical models alone [ 39 40 ]. Additionally, image-based prognostic signatures may help predict survival outcomes and guide surgical decisions more accurately than conventional radiological criteria, thus enhancing the stratification of potential surgical candidates [ 41 42 ]. Ongoing clinical trials are essential to validate the clinical applicability of these emerging imaging technologies. Integrating AI and radiomics into clinical practice could improve surgical outcomes, reduce recurrence, and ultimately enhance survival through more tailored treatment strategies.

Peritoneal disease is a hallmark of advanced OC and is present in most patients at diagnosis. CE-CT is primarily used to evaluate the site and extent of tumor metastasis. However, CE-CT tends to underestimate peritoneal metastases ( Fig. 1 ), especially in the gastrointestinal ( Figs. 2 , 3 , 4 ) and mesenteric regions, while demonstrating relatively high accuracy for diaphragmatic and omental involvement ( Fig. 5 ). Consequently, CE-CT cannot reliably predict suboptimal cytoreduction, with reported 79% sensitivity and 75% specificity [ 15 ]. In contrast, mp-MRI offers a distinct advantage with the ability to employ functional imaging techniques such as DWI to detect and quantify small peritoneal metastases ( Fig. 1 ). The peritoneal cancer index (PCI) is a scoring system developed to evaluate the extent of peritoneal carcinomatosis based on the size and distribution of tumor lesions within the peritoneal cavity [ 16 ], which can be conducted either surgically or radiologically [ 17 ]. Studies indicate that DWI-MRI can reliably assess the extent of peritoneal metastases using the PCI system [ 18 19 ]. Whole-body MRI (WB-MRI) has been shown to categorize tumor volume more accurately than CE-CT (91% vs. 50%), with WB-MRI showing superior sensitivity (95% vs. 55%) and accuracy (88% vs. 63%) according to the tumor site. Moreover, other studies corroborated the accuracy of MRI (36) in determining tumor extent, with findings comparable to the surgical PCI score (33) [ 20 ]. The superior performance of mp-MRI in peritoneal disease staging was further substantiated in a meta-analysis (pooled sensitivity, 92%; specificity, 85%) compared with CE-CT and PET-CT [ 21 ]. Additionally, the ability of mp-MRI to predict suboptimal cytoreduction is well-established, with a sensitivity of 94.0% and specificity of 97.7% [ 22 ]. A multicenter study, MRI in OC (MROC), prospectively validated the accuracy of mp-MRI compared with CE-CT scans for classifying and delineating the extent of OC [ 23 ]. The results are eagerly anticipated and have the potential to highlight the superior diagnostic capabilities of mp-MRI, which are essential in improving staging and surgical planning. The role of 18 F-fluorodeoxyglucose (FDG)-PET-CT in initial staging remains inconclusive. Some studies suggest its potential role in OC metastatic disease detection, monitoring treatment response, and identification of recurrence [ 21 ]. However, it is associated with high costs and has limited ability to detect discrete/low-volume peritoneal tumors, hindering accurate staging. Lopez-Lopez et al. [ 24 ] demonstrated CE-CT's superior diagnostic accuracy compared to 18 F-FDG-PET-CT for confirming peritoneal disease, while also highlighting the potential of the latter for identifying extraperitoneal metastases.

Recurrence is common in advanced OC, often presenting in the peritoneum, pelvis, or distant sites [ 7 ]. Detecting recurrences can be difficult, as the features may be subtle to distinguish from post-treatment changes such as fibrosis and inflammation. Serum CA125 in conjunction with serial CE-CT remains the most widely used surveillance imaging for recurrence, although its sensitivity and specificity are limited, especially in discrete lesions (≤1 cm) [ 35 ]. In recent years, FDG-PET-CT has emerged as an invaluable tool in detecting OC recurrence, combining anatomical characterization with metabolic information. Studies have shown that FDG-PET-CT has high sensitivity (91%) and specificity (88%), outperforming CE-CT and mp-MRI, particularly when findings are inconclusive [ 36 ]. A key role is in patients with increasing CA125, with no clear site of disease on CE-CT. LNs are the most common site of OC relapse [ 37 ]. FDG-PET-CT is particularly effective in detecting metastatic LN (i.e. hypermetabolic activity), whereas CE-CT may only show nonspecific changes ( Fig. 16 ) [ 35 ]. Additionally, benign post-treatment changes do not typically exhibit FDG uptake, aiding differentiation from malignant recurrences [ 37 ]. The value of maximal cytoreduction in both primary and relapsed OC has been primarily demonstrated in patients without preoperative PET-CT, focusing on removing macroscopic disease. However, PET-CT's role in detecting non-bulky microscopic diseases to exclude patients from macroscopic cytoreduction has not been evaluated in clinical trials. The LION study showed that removing microscopically involved LN did not impact survival, suggesting that similar findings may apply to other microscopically affected sites [ 29 ]. Therefore, PET-CT results should be carefully interpreted along with clinical symptoms, treatment options, and conventional imaging.

Imaging is integral to the diagnosis, staging, and treatment planning of advanced OC. While CT remains central, MRI has become an integral part of diagnostic investigation due to its superior accuracy in the detection and extent of peritoneal disease and characterization of adnexal masses. FDG-PET-CT may be helpful in the recurrent disease setting and in evaluating treatment responses. Radiomics is an emerging area of research with great potential. Integration of standardized imaging protocols and lexicons with advanced imaging technologies will further improve diagnostic capabilities, refine clinical decision-making, and drive personalized OC care.

Despite advances, challenges persist in the imaging of advanced OC. Detecting small peritoneal metastases, particularly in the mesentery and gastrointestinal tract, remains difficult despite the use of high-resolution CE-CT and mp-MRI. Furthermore, distinguishing between malignant and benign conditions, such as endometriosis and degenerating fibroids, can lead to diagnostic pitfalls [ 38 ]. Accurate staging of LN involvement, especially retroperitoneal or supradiaphragmatic lymphadenopathy, may require additional imaging techniques or biopsy where clinically indicated [ 39 ]. These limitations emphasize the importance of multimodal imaging approaches, where the combination of CE-CT, mp-MRI, and PET-CT enhances overall diagnostic accuracy. Furthermore, false positives and false negatives, as well as the lack of standardized imaging protocols, highlight the necessity for further validation through multicenter trials to ensure consistent and reliable results across diverse clinical settings.

In advanced OC, parenchymal metastasis to solid abdominal organs, including the liver ( Figs. 6 , 7 , 8 , 9 , 10 ) and spleen ( Fig. 11 ), is common and carries significant clinical implications (i.e., prognosis and treatment strategies). CE-CT has high diagnostic accuracy for detecting metastatic lesions in the abdominal parenchyma, with similar sensitivity ( P = 0.290) and lower specificity compared to PET-CT ( P < 0.010) [ 25 ]. However, its sensitivity for smaller lesions and those located in complex anatomical regions such as the mesentery or retroperitoneum remains limited. MRI, specifically DWI, can help clarify indeterminate parenchymal and liver deposits detected on CE-CT, providing higher conspicuity for the detection of malignant deposits [ 26 ].