Preoperative Magnetic Resonance Imaging versus Intraoperative Frozen Section Diagnosis for Predicting the Deep Myometrial Invasion in Endometrial Cancer: Our Experience and Literature Review

Multi-modality MRI radiomics phenotypes in intermediate-high risk endometrial cancer: correlations with histopathology and prognosis

OBJECTIVES

This study aimed to identify the magnetic resonance imaging (MRI)-based radiomics phenotypes of intermediate-to-high-risk endometrial cancers (ECs), explore their association with histopathologic features, and compare their prognostic ability with the International Federation of Gynecology and Obstetrics (FIGO) stage.

METHODS

This study retrospectively recruited 355 patients with pathologically confirmed EC from 01/2016 to 06/2023. 166(46.8%) were classified as intermediate-to-high-risk ECs according to the European Society for Medical Oncology guidelines. Radiomics clustering analysis was performed on preoperative MRI to identify the radiomics phenotype of intermediate-to-high-risk ECs. The association between the radiomics phenotypes and the clinicopathologic information was explored, and the added value in predicting the recurrence was also evaluated using concordance index (C-index).

RESULTS

Of the included 166 patients (average age 56.83 ± 9.25 years), 23 were recurrent patients. The corresponding tumors in various clusters were assigned to phenotypes 1 and 2. Larger tumor diameter (P < .01), cervical mucosa invasion [30(36.15%) vs 15(18.07%), P = .01], deep myometrial infiltration [51(61.45%) vs 31(37.35%), P = .00], and histologic subtype [17(20.48%) vs 5(6.02%), P = .01] were associated with subtype 1. The risk of recurrence (P = .01) was higher in phenotype 1, and the FIGO stage could further differentiate higher recurrence risk in phenotype 1 (P < .01). The C-index was 0.66 for the radiomics phenotype model, 0.69 for the FIGO stage model, and 0.72 for the combined model.

CONCLUSIONS

MRI-based radiomics consensus clustering enabled the identification of associations between radiomics features and histopathologic features in intermediate-to-high-risk EC. The FIGO stage could further elevate the prediction ability of recurrence risk.

Ultrasonography, macroscopy, and frozen section: whıch is better for predicting deep myometrial invasıon in endometrial cancer?

OBJECTIVE

The aim of this study was to compare the power of preoperative transvaginal ultrasonography, intraoperative macroscopic examination, and frozen section for predicting deep myometrial invasion in endometrial cancer.

METHODS

This is a retrospective review involving 68 patients who underwent surgical staging for endometrial cancer from 2014 to 2017. Patients with grade 3 endometrial cancer and non-endometrioid tumors were excluded. The findings related to preoperative transvaginal ultrasonography, intraoperative macroscopic examination, and frozen section were compared with definitive histopathological diagnosis.

RESULTS

The mean age, gravidity, and body mass index of the patients were 58.1±8.9 years (range: 30-80 years), 3.2±2.1 (range: 0-9), and 33.5±6.6 kg/m2 (range: 20-52 kg/m2), respectively. Only 11 (16.2%) patients were in the premenopausal period, while 57 (83.8%) were in the postmenopausal period. Grade 1 endometrial cancer was found in 29 patients (42.6%) and grade 2 tumors were specified in 39 patients (57.4%). Stage IA disease was found in 45 (66.2%) patients, while stage IB disease was observed in 23 (33.8%) patients. The 5-year survival rate was 91.2%. The sensitivity of preoperative transvaginal ultrasonography, intraoperative macroscopic examination, and frozen section were 56, 34, and 52%, respectively, for predicting deep myometrial invasion. In contrast, the specificity of preoperative ultrasonography, intraoperative macroscopic examination, and frozen section were 86, 100, and 100%, respectively.

CONCLUSION

Transvaginal ultrasonography and intraoperative frozen section were found to have similar sensitivity and specificity for predicting deep myometrial invasion. Preoperative transvaginal ultrasonography appears as an efficient approach for predicting endometrial cancers with deep myometrial invasion.

Evaluation of the role of intraoperative frozen section and magnetic resonance imaging in endometrial cancer

OBJECTIVE

To evaluate the accuracy of preoperative endometrial biopsy and magnetic resonance imaging (MRI) of endometrial cancer compared with that of intraoperative frozen section.

METHODS

This retrospective study included 264 patients who underwent surgery with intraoperative frozen section for endometrial cancer at our institution between 2014 and 2018. Diagnosis was determined by histologic type, grade, and myometrial invasion. Concordance rate, sensitivity, and specificity of preoperative diagnosis and intraoperative frozen diagnosis were calculated, in comparison to the final pathologic diagnosis.

RESULTS

Preoperative and intraoperative diagnoses showed no statistically significant difference in determining histologic type and grade (P = 0.152). Intraoperative diagnosis showed higher sensitivity for endometrioid carcinoma grade 3 and other types, and higher specificity for grade 1. For myometrial invasion, intraoperative diagnosis showed significantly higher concordance rate than preoperative MRI findings (P < 0.01). Intraoperative diagnosis showed higher sensitivity and specificity in patients with and without myometrial invasion, respectively.

CONCLUSION

Higher agreement between intraoperative and final diagnoses, especially in myometrial invasion, suggests that intraoperative frozen section is a good indicator for appropriate surgical procedure decision making.

Analysis of vanishing endometrial cancer by pathological types

PURPOSE

We asked why endometrial cancer sometimes vanishes.

METHODS

A total of 454 patients diagnosed with endometrioid-type endometrial cancer (EC) (via endometrial sampling) and treated in our clinic over the past 5 years were enrolled. The patients were divided into two groups: vanishing and residual, depending on whether a tumor was detected in the postoperative hysterectomy specimen. Patient age, numbers of pregnancies and deliveries, menopausal status, systemic disease status, hemogram parameters, International Federation of Gynecology and Obstetrics (FIGO) grade, and invasion status (evident on magnetic resonance imaging [MRI]) were compared between the groups.

RESULTS

ECs vanished in 42 (9.25%) patients. The vanishing rates were 19.7% (37/187) in FIGO grade 1 patients, 2.1% (5/238) in grade 2 patients, and 0% (0/29) in grade 3 patients. The average age was lower in the vanishing than the residual group, but the premenopausal status and grade 1 tumor rates were higher (both p < 0.001). An absence of invasion (as revealed by MRI) was more common in the vanishing group (p < 0.001). No recurrence developed in the vanishing group, but recurrences were noted in 3.3% (14/412) of the residual group. There were no significant between-group differences in any of the numbers of pregnancies or births, systemic disease status, or hemogram parameters (all p > 0.05).

CONCLUSION

Vanishing EC is more likely in premenopausal women with endometrioid grade 1 EC (as revealed by endometrial biopsy) who lack myometrial invasion on MRI.

Efficacy of transvaginal ultrasound versus magnetic resonance imaging for preoperative assessment of myometrial invasion in patients with endometrioid endometrial cancer: a prospective comparative study

Background

We compared the accuracy of preoperative transvaginal ultrasound (TVUS) versus magnetic resonance imaging (MRI) for the assessment of myometrial invasion (MI) in patients with endometrial cancer (EC), while definitive histopathological diagnosis served as a reference method.

Patients and methods

Study performed at a single tertiary centre from 2019 to 2021, included women with a histopathological proven EC, hospitalized for scheduled surgery. TVUS and MRI were performed prior to surgical staging for assessment MI, which was estimated using two objective TVUS methods (Gordon’s and Karlsson’s) and MRI. Patients were divided into two groups, after surgery and histopathological assessment of MI: superficial (≤ 50%) and deep (> 50%).

Results

Sixty patients were eligible for the study. According to the reference method, there were 34 (56.7%) cases in the study with MI < 50%, and 26 (43.3%) with MI > 50%. Both objective TVUS methods and MRI showed no statistical significant differences in overall diagnostic performance for the preoperative assessment of MI. The concordance coefficient between both TVUS methods, MRI and histopathology was statistically significant (p < 0.001). Gordon’s method calculating MI reached a positive predictive value (PPV) of 83%, negative predictive value (NPV) of 83%, 77% sensitivity, 88% specificity, and 83% overall accuracy. Karlsson’s method reached PPV of 82%, NPV of 79%, 69% sensitivity, 88% specificity, and 80% overall accuracy. Accordingly, MRI calculating MI reached PPV of 83%, NPV of 97%, 97% sensitivity, 85% specificity, and 90% overall accuracy.

Conclusions

We found that objective TVUS assessment of myometrial invasion was performed with a diagnostic accuracy comparable to that of MRI in women with endometrial cancer.