{"paper_id":"c8f2a9e7-e896-4d41-905b-3547067a8020","body_text":"Endometrial cancer (EC) is currently the sixth most common cancer among women and the eleventh most common cause of cancer-associated death in women worldwide. 1  The incidence of EC is rising, and by 2040, it is projected to become the third most common cancer and the fourth leading cause of cancer-related deaths in women. Notably, the increase in incidence is particularly pronounced among non-Hispanic Black, Hispanic, and Asian women, who tend to develop more aggressive forms of the disease. Alarming trends in mortality rates are likely driven by a higher prevalence of advanced-stage cancers, high-risk histologic types, and diagnoses occurring after the age of 65. The current median age at EC diagnosis is 63, with around 80% of cases occurring in women aged 55 years and older. However, a growing number of cases are being observed in premenopausal women, for whom fertility-preserving treatments may be especially critical. 2\nMost risk factors are modifiable lifestyle-related factors, including a high-fat diet and lack of regular physical activity. Other key risk factors include increasing age, obesity, and prolonged estrogen exposure—including early menarche, late menopause, nulliparity, and the use of tamoxifen. Genetically, EC is linked to Lynch syndrome, an autosomal dominant disorder caused by germline defects in mismatch repair genes ( MLH1 ,  MSH2 ,  MSH6 , and  PMS2 ). Individuals with this condition face an 80% lifetime risk of developing colorectal cancer and a 60% risk of EC. 3\nEC is usually diagnosed at an early stage, with postmenopausal uterine bleeding being the most common presenting symptom. In premenopausal or perimenopausal women, this symptom can sometimes be misattributed to expected irregular menstrual cycles due to anovulatory conditions like polycystic ovary syndrome. 3\nThe initial diagnostic approach involves a physical examination, followed by transvaginal ultrasound (US), which has a high negative predictive value, and endometrial biopsy (via hysteroscopy or dilatation and curettage). For staging, pelvic MRI is preferred due to its superior soft tissue resolution, offering detailed anatomical insights. 3\nOver the past decade, considerable advances have been made in the management of EC, driven by improved molecular classifications and the emergence of new therapeutic options. This manuscript will delve into the latest trends in histo-molecular and clinical management of EC, highlighting the central role of MRI, typical radiologic findings, and emerging perspectives on the integration of imaging throughout a patient’s treatment journey. 3\n\nEC is highly heterogeneous, both at the histologic and molecular level. Initially, EC was defined histologically as either type 1 or type 2 disease: type 1 consisted of low-grade, endocrine receptor–positive endometrioid cancers (around 60% of EC), and type 2 encompassed high-grade endometrioid and nonendometrioid subtypes (i.e., serous, clear cell, undifferentiated carcinoma, carcinosarcoma, and mixed). 4  In 2013, a major publication from The Cancer Genome Atlas led to the creation of molecularly distinct subgroups associated with specific prognoses and directed therapies. 4  The subgroups include 1)  POLE /ultra-mutated genetic profiles (≈10%), which exhibit excellent prognosis and an extremely low recurrence rate; 2) dMMR/mismatch repair deficiency (≈30%), which has an intermediate prognosis due to the enhanced sensitivity to immunotherapy; 3)  TP53  mutations (≈20%), which have a low mutational load and genome instability and are associated with the poorest prognosis; and 4) NSMP/no specific molecular profile (≈40%), which has an intermediate prognosis and frequent expression of estrogen and progesterone receptors. 4\nAs introduced earlier, the initial treatment of EC is surgical, but adjuvant treatment depends on histopathologic and molecular features. The phase 3 PORTEC-3 trial was conducted to assess whether the addition of adjuvant chemotherapy during and after radiotherapy could improve outcomes for women with high-risk EC compared to pelvic radiotherapy alone. Following the International Federation of Gynecology and Obstetrics (FIGO) staging 2009, patients at different stages of EC were included in the trial: high-risk EC at stage I, endometrioid-type grade 3 with deep myometrial invasion or lymphovascular space invasion, endometrioid-type EC at stage II or III, or EC at stage I to III with serous or clear cell histology. Results demonstrated improved survival with the addition of chemoradiotherapy. 5  However, an ancillary study revealed that molecular profiling significantly influenced 5 year recurrence-free survival rates, with a recurrence-free survival rate of 48% for  TP53 -mutated tumors, 98% for  POLE -mutated, 72% for dMMR, and 74% for NSMP tumors ( P  < 0.001). These findings contributed to the 2023 update in EC staging (FIGO 2023), 6  which now incorporates molecular characteristics alongside traditional histologic and cancer staging criteria. 5 , 7\nThe critical updates of the FIGO 2023 classification ( Fig. 1 ) 8  include:\n Refined classification : The system now accommodates the diverse histologic and molecular profiles seen in EC, capturing the complexity of its subtypes. Molecular marker integration : New prognostic insights are derived from molecular markers, aligning EC staging methods more closely with breast cancer practices. Advanced risk stratification : Leveraging protocols from prominent oncologic societies, the revised staging emphasizes molecular aggressiveness over disease extent, enhancing prognostic accuracy and treatment planning.\nRefined classification : The system now accommodates the diverse histologic and molecular profiles seen in EC, capturing the complexity of its subtypes.\nMolecular marker integration : New prognostic insights are derived from molecular markers, aligning EC staging methods more closely with breast cancer practices.\nAdvanced risk stratification : Leveraging protocols from prominent oncologic societies, the revised staging emphasizes molecular aggressiveness over disease extent, enhancing prognostic accuracy and treatment planning.\n\nAlthough MRI is not part of the FIGO 2023 update, it does have a role in helping to distinguish EC from cervical cancer, treatment planning, particularly in deciding the extent of surgery and a patient’s candidacy for fertility-preserving options, radiotherapy planning, and follow-up. The main revisions related to imaging include 1) assessing for the presence of myometrial invasion, regardless of whether it is more or less than 50%; 2) evaluating synchronous and metastatic ovarian tumors; and 3) determining peritoneal involvement. 9\nThe European Society of Urogenital Radiology–Society of Abdominal Radiology (ESUR-SAR) guidelines 10  are widely recognized internationally; other guidelines, including those by the Japan Radiological Society (JRS), are designed to accommodate specific national circumstances. 11  The protocol should include at least the following:\n High-resolution small FOV T2-weighted imaging (T2WI), diffusion-weighted images (DWIs) with b-values of at least 0 and at least 800 sec/mm 2 ; T2WI and DWI should be acquired both in the sagittal plane and axial-oblique plane perpendicular to the endometrial cavity, i.e., along the short axis of the uterine body ( Fig. 2.1 ). It helps to evaluate for the presence of myometrial invasion, tumor local extent, and nodal metastases. Dynamic contrast-enhanced (DCE) images should be obtained in the sagittal plane, and delayed DCE should be obtained in the axial-oblique plane to assess different involvement extents ( Fig. 2.2 ). The temporal resolution should be 30s or less for a minimum of 2 min.\nHigh-resolution small FOV T2-weighted imaging (T2WI), diffusion-weighted images (DWIs) with b-values of at least 0 and at least 800 sec/mm 2 ; T2WI and DWI should be acquired both in the sagittal plane and axial-oblique plane perpendicular to the endometrial cavity, i.e., along the short axis of the uterine body ( Fig. 2.1 ). It helps to evaluate for the presence of myometrial invasion, tumor local extent, and nodal metastases.\nDynamic contrast-enhanced (DCE) images should be obtained in the sagittal plane, and delayed DCE should be obtained in the axial-oblique plane to assess different involvement extents ( Fig. 2.2 ). The temporal resolution should be 30s or less for a minimum of 2 min.\nThe early postcontrast phase at 30 to 60s after IV contrast administration in the sagittal plane is best to evaluate for superficial myometrial invasion. Uninterrupted enhancement of the subendometrium or inner junctional zone, which usually enhances earlier than the myometrium, excludes superficial invasion. The equilibrium phase at 120 to 180s can be obtained as a part of DCE or single-phase axial-oblique acquisition to evaluate for deep myometrial invasion by contrasting the hypointense tumor with the enhancing background myometrium. Finally, the delayed phase at 4 to 5 min in the sagittal and/or axial-oblique plane helps evaluate for cervical stromal invasion defined by the interruption of the endocervical mucosa. 12 – 15  These considerations along with the normal anatomy are reflected on  Fig. 2 .\nThe difference between ESUR/SAR and JRS MRI protocol guidelines arises in using the large FOV T1-weighted imaging (T1WI), with or without T2WI, to evaluate for nodal and distant metastases. While ESUR/SAR recommends including the kidneys to the iliac crests, JRS recommends only covering the latter when CT is omitted. 10 , 11\n\nAs introduced earlier, endometrial sampling is recommended for further evaluation in symptomatic postmenopausal women with an endometrial thickness of 5 mm or more on transvaginal US ( Fig. 3 ). However, if the endometrium is thin in US but the patient has persistent or recurrent uterine bleeding, hysteroscopy should be performed to evaluate for high-risk histologies. 15  In contrast, the referral guidelines, especially regarding the endometrial thickness threshold, are less clear for asymptomatic and premenopausal women, and gynecologists make management decisions based on clinical assessment and a review of risk factors. Similarly, when unsuspected endometrial thickening is identified on a CT scan performed for other reasons, a transvaginal US is advised for further evaluation. This is because factors like small hydrometra, hematometra, and menstrual cycle changes can affect the appearance of the endometrium and mimic pathologic thickening. 15\nThe classic appearance of EC includes an intermediate-to-high signal intensity on T2WI, contrasting with the normal hyperintense endometrium and hypointense junctional zone. The tumor usually demonstrates diffusion restriction on DWI, i.e., high-signal intensity on high b-value DWI and low-signal intensity on the corresponding apparent diffusion coefficient (ADC) map. Finally, it typically enhances homogeneously and less intensely than the adjacent myometrium, with maximal endometrium-to-myometrium contrast at 90 to 120s following IV contrast administration. 13\nDWI adds considerable value, especially in early-stage cases where small tumors may not cause noticeable endometrial thickening and in cases where the endometrial cavity is nondistended. Multiple studies have confirmed the utility of DWI in distinguishing EC from normal endometrium, with ADC values of ECs substantially lower than those of endometrial polyps and normal tissue. ADC thresholds ranging from 1.05 to 1.28 mm 2 /sec have been proposed to differentiate malignant from benign lesions, with sensitivity ranging from 60.1% to 87% and specificity reaching 100%. This is particularly relevant in cases where endometrial sampling is not feasible. 16 , 17  However, false-positive cases of malignancy on DWI can arise due to physiologic secretory and hyperplastic endometrium, which have low ADC values. Furthermore, blood product retention can also demonstrate low ADC values. However, the T1WI or postcontrast subtraction sequences can help confirm the presence of blood products. 2 , 10 , 18\nThe goal of MRI in FIGO 2023 stage I is to assess whether myometrial invasion is present. While FIGO 2009 defined the stage based only on myometrial invasion, the FIGO 2023 considers the histologic type and degree of lymphovascular invasion (LVSI). Therefore, the new substages include 1) stage IA1, with nonaggressive histologies limited to an endometrial polyp or confined to the endometrium; 2) stage IA2, with nonaggressive histology but myometrial involvement <50% and negative or focal LVSI; 3) IA3, low-grade endometroid carcinoma, absence of extensive or substantial LVSI, and myometrial involvement <50% with unilateral ovarian involvement without capsule invasion or rupture; 4) IB, with nonaggressive histology and negative or focal LVSI but myometrial involvement ≥50%; and 5) IC, aggressive histology limited to a polyp or without myometrial invasion. 3\nMRI assessment for myometrial invasion is best achieved using a side-by-side review of T2WI, DWI, and DCE. MRI accurately depicts myometrial invasion and its depth, with a reported sensitivity and specificity of 86% for deep myometrial invasion 19  and 91.1% sensitivity and 46.1% specificity for detecting any myometrial invasion. 20  The criteria to rule out myometrial invasion includes a continuous/uninterrupted endometrial-internal myometrial junction both on T2WI/DWI and subendometrial enhancement, as well as the presence of a thin band of avid enhancement between the endometrium and myometrium less than 60s after IV contrast administration on DCE 10 , 21  ( Fig. 2 ).\nTo assess the depth of myometrial invasion, a line should be drawn along the anticipated inner edge of the endometrium-myometrium junction on axial-oblique images. Next, 2 additional lines are drawn: one to measure the overall thickness of the myometrium and the other to measure the maximum extent of tumor invasion within the myometrium. The ratio of these 2 measurements represents the percentage of myometrial invasion 10 , 16  ( Fig. 4.1  and  4.2 ). Interestingly, reduced FOV DWI has shown increased diagnostic accuracy when compared to T2WI and DCE MRI (accuracy 84%–96% vs 92%–98%). 22\nLow-signal-intensity tumors within the outer myometrium or beyond indicate deep myometrial invasion. Furthermore, disruption of the low-signal-intensity contour of the uterine serosa and/or an irregular uterine contour on T2WI and loss of the regular rim of highly enhancing myometrium on DCE indicate serosal involvement 10 , 16 , 21  ( Fig. 5.1  and  5.2 ).\nSeveral coexisting conditions can challenge the assessment of myometrial invasion. First, evaluation in cases with coexisting adenomyosis can result in overestimation of invasion due to interruption of subendometrial enhancement by adenomyosis and the increase in thickness of the junctional zone ( Fig. 4 ). Second, leiomyomas can distort the uterine zonal anatomy from mass effect. Third, myometrial thinning in the case of large tumors (which can distend the endometrial cavity and compress the endometrium), or physiologic thinning of the myometrium found in the cornual regions can occur. The deep myometrial invasion should not be overcalled in these cases ( Figs. 6.1  and  6.2 ). Finally, an iso-intense tumor on T2WI can be challenging to distinguish from the normal myometrium. 9  In such cases, careful lesion evaluation on the 3 main sequences and recognition of the typical imaging appearances of the associated conditions (for instance, adenomyosis does not demonstrate diffusion restriction and enhances variably) can help overcome these challenges. 14\nAll of the features described are of paramount importance when deciding whether a more conservative management approach is possible: 1) If the patient desires to avoid a hysterectomy to preserve fertility and opts for treatment with hormonal therapy, then myometrial invasion must be absent on both MRI and histology after hysteroscopy or dilatation and curettage. 2) If a patient with EC desires ovarian preservation, this could be considered in premenopausal patients with low-risk histology, myometrial invasion of less than 50%, and no extrauterine disease. 23\nThe new FIGO 2023 has incorporated more factors than the cervical stromal invasion into the definition of stage II EC, including: 1) IIA, nonaggressive histology but with cervical stromal invasion; 2) IIB, also nonaggressive histology but with extensive/substantial LVSI and any myometrial invasion; and 3) IIIC, aggressive histology with any myometrial invasion. Therefore, it is essential to be aware of presurgical biopsy results to stage EC on imaging properly. 3  Of these features, cervical stromal invasion is the only one that can be reliably evaluated on MRI. It is associated with well-established adverse prognostic factors such as LVSI or lymph node metastases; therefore, it is a poor prognostic factor for progression-free and overall survival. 24  LVSI is only evaluable on histology, but there are a number of articles published evaluating radiomics models to attempt to predict its presence on MRI. 25  Most of them evaluated a radiomic model (analyzing for example: tumor shape, texture, and intensity) combined with clinical and MRI features to predict preoperatively LVSI. 25\nTo suspect cervical invasion on imaging, the tumor must disrupt the normal cervical stroma and endocervical canal. In premenopausal patients, the inner cervical stroma (low-to-intermediate T2-signal intensity) is distinguished from the outer cervical stroma (intermediate-to-high signal intensity). Following contrast administration, the endocervical mucosa typically shows rapid, avid, and persistent enhancement. At the same time, the cervical stroma enhances gradually on DCE, becoming more conspicuous in the delayed phases (≥4 min) ( Fig. 7 ). 10  Himoto et al 26  found that the combination of T2WI, DWI, and DCE outperformed T2WI and DWI for cervical stroma assessment (area under the curve [AUC] 0.99 [sensitivity, 0.01; 95% confidence interval (CI), 0.91–1.00] vs AUC 0.79 [95% CI, 0.64–0.89], respectively).\nSome pitfalls should be noted when assessing EC on MRI. A tumor protruding into the internal os does not necessarily signify cervical stromal invasion; invasion can only be diagnosed when the cervical stroma is disrupted ( Fig. 7 ). Additionally, cervical stromal invasion may occur without involvement of the endocervical mucosa, particularly due to adjacent myometrial invasion. 9  Furthermore, Nabothian cysts in the cervix can cause localized areas of high signal intensity on DWI due to the T2WI shine-through effect. In such cases, the absence of a corresponding low signal on the ADC map helps differentiate the cysts from tumor extension within the cervix. 27\nAs mentioned earlier, discriminating cervical carcinoma from stage II EC is challenging ( Fig. 8 ). Some helpful MRI clues are based on identifying the tumor epicenter and knowledge of typical disease spread and enhancement patterns. Cervical adenocarcinomas are more likely to be centered in the cervical region, with possible parametrial invasion, and are more frequently hypervascular. Contrary to this, EC is centered in the endometrial cavity, involving endometrium and myometrium, and is more regularly hypoenhancing. 18\nAll the factors mentioned above impact the treatment decision-making process; if fertility preservation is being considered, then the presence of cervical stromal invasion excludes it as an option and also helps define the type of hysterectomy required. Whereas a simple hysterectomy involves resection of the uterus and is a less extensive surgery, patients with cervical stromal invasion may be considered for a radical hysterectomy, consisting of resection of the uterus, cervix, and parametrium, which is a more extensive surgery with more potential complications. However, most studies have shown that a simple hysterectomy is still the recommendation for any EC diagnosis. 3\nStage III now differentiates between adnexal and uterine serosal infiltration and includes refinements for lymph node metastases; including the pelvic and para-aortic nodes to the level of the renal vessels. Nodal involvement is also subcategorized by micro- and macrometastases. 3  It includes the following substages according to the anatomic site involved, without factoring in the histologic subtype: 1) IIIA1, ovaries and fallopian tubes; 2) IIIA2, subserosa and serosa; 3) IIIB1, vagina and parametrium; 4) IIIB2, pelvic peritoneum; 5) IIIC1i, pelvic node micrometastasis; 6) IIIC1ii, pelvic node macrometastasis; 7) IIIC2i, infrarenal node micrometastasis; and 8) IIIC2ii, infrarenal node macrometastasis.\nThis should be suspected when there is disruption of the T2WI hypointense signal of the uterine serosa, irregularity of the outer uterine contour, and interruption of the typical rim enhancement of the outer myometrium on DCE. 14\nEC has up to a 30% risk of synchronous primary or metastatic ovarian cancer. The risk is higher among patients under 45 years, adversely impacting oncologic outcomes. 3  Previously, it was deemed essential to differentiate between EC with ovarian metastasis and contemporaneous EC and ovarian cancer ( Fig. 9 ). While the former is the expected scenario in patients with high-grade EC, the latter could occur predominantly in low-grade EC, in which some clonal link exists between EC and ovarian cancer. 28 , 29\nIn general, evaluation of the adnexa should follow the recommendations of the OvarianAdnexal Reporting and Data System (O-RADS) MRI risk stratification system, which has been described extensively in the literature. 30  However, a few signs could help differentiate between synchronous malignancies and metastasis. The signs favoring synchronous ovarian and ECs include early-stage EC with minimal or no myometrial invasion and a unilateral mass on a background of endometriosis or a borderline tumor. On the other hand, features favoring ovarian metastasis from EC include an endometrial mass with deep myometrial invasion and/or tubal invasion and bilateral ovarian masses. DWI can also detect adnexa deposits, especially in high-risk histologies. 10 , 21 , 31\nThe most common lymphatic chains involved in EC are the pelvic and para-aortic regions. 3  MRI evaluation relies on short-axis diameter enlargement (pelvic lymph nodes >8 mm and abdominal lymph nodes >10 mm) or abnormal morphology (irregular contours, heterogeneous signal intensity, rounded, or necrotic). However, if there is suspicion of nodal involvement, PET/CT is recommended to provide a more accurate extent of the disease. 14\nOn DWI, the lymph nodes can be easily detected as high signal-intensity ovoid structures. Although attempts have been made to distinguish between malignant and benign lymph nodes based on ADC values, most studies found a certain degree of overlap, and none could define the negative predictive value of DWI. 32\nThis stage encompasses disease spread to the mucosa of the bladder or bowel, or the presence of distant metastasis. It is divided into 1) IVA, involvement of the mucosa of the bladder and/or rectum; 2) IVB, peritoneal metastasis outside the pelvis; and 3) IVC, suprarenal/inguinal lymph node involvement, and/or distant metastasis.\nStage IVA, representing bladder or rectal mucosal involvement, is best evaluated on MRI of the sagittal plane. On T2WI, it is diagnosed when the tumor abuts or indents the bladder/rectum over a marked surface area, interrupts the low-signal intensity of the bladder/rectal muscular layer, invades the bladder/rectal muscular wall, and when tumor nodules are detected in the mucosal layer. However, bullous edema of the bladder may be a sign of a tumor in the subserosal or muscular layer of the bladder but not in the mucosa. 33\nLiu et al. 34  conducted a retrospective study that included 2948 patients with stage IVC EC.\nThe most common metastatic sites included the lungs (37.0%), liver (11.9%), bones (6.8%), and brain (2.0%). Patients with metastases at each site had a different median survival: 10 months for lung, 8 months for liver, 6 months for bone, and 5 months for brain metastasis ( P  = 0.003). 34\nCT is routinely used to assess metastatic disease, and  18 F-fluorodeoxyglucose positron emission tomography ( 18 F-FDG PET)/CT helps evaluate nodal and distant metastases ( Fig. 10 ). Some institutions have implemented  18 F-FDG PET/MRI as a “one-stop shop” modality to assess for local recurrent disease and distant metastasis. 35 – 39  The systematic review performed by Bollineni et al. 40  (n = 13) found that FDG PET/CT had pooled sensitivity of 0.72 (95% CI, 0.63–0.80) and specificity of 0.94 (95% CI, 0.93–0.96), with an AUC of 0.94 and a Q* value of 0.88 for detecting lymph node metastasis. For detecting recurrent endometrial cancer, the pooled sensitivity was 0.95 (95% CI, 0.91–0.98) and the pooled specificity was 0.91 (95% CI, 0.86–0.94), with an AUC of 0.97 and Q* values of 0.93. 40\n\nMost cases of tumor recurrence occur within the first 2 years of treatment, with the most common sites being the vaginal vault and lymph nodes, followed by metastatic disease to the liver, lungs, peritoneum, and bones. 14\nFor pelvic recurrence, MRI is the modality of choice for evaluation ( Figs. 11  and  12 ). The expected postoperative findings include a smooth vaginal vault without new solid nodules, complex cystic masses, or diffuse plaque-like thickening with restricted diffusion and/or abnormal enhancement. The anticipated post-radiation findings include decreased uterine volume, endometrial thinning, loss of standard uterine zonal anatomy definition with diffuse T2WI-hypointense signal, lack of restricted diffusion, and delayed enhancement. The most notable pitfall in these cases is the discrimination of recurrence from posttreatment changes (including edema), for which appropriate timing of follow-up imaging after treatment completion (at least 3–6 months), and/or biopsy could greatly help. 14\n\nThere are a few published templates for structured reporting of EC on imaging. 10 , 21 , 41  The essential elements of the report include 1) tumor size; 2) presence and depth of myometrial invasion; 3) presence of cervical stromal invasion; 4) invasion of the vagina, adnexa, and uterine serosa; 5) invasion of the bladder, rectum, or pelvic wall; 6) metastasis to pelvic, inguinal, and para-aortic lymph nodes; 7) presence of hydronephrosis; and 8) presence of distant metastasis.\n\nThe prevalence of EC is increasing, and dedicated MRI is clinically beneficial in initial staging, determining appropriate treatment options balanced with patient wishes and surveillance for recurrence. For clinicians in daily practice, knowledge of key clinical features, imaging appearances and staging features of EC, and the expected posttreatment changes are paramount to providing high-level patient care.","source_license":"CC-BY-4.0","license_restricted":false}