Evaluating equity of access and predictors of minimally invasive hysterectomy for endometrial and cervical cancer from 2000 to 2017 in Ontario, Canada: A population-based cohort study.

INTRODUCTION: We sought to assess the uptake of minimally invasive hysterectomy among patients with endometrial and cervical cancer in Ontario, Canada, and assess the equity of access to minimally invasive surgery (MIS) by evaluating associations with patient, disease, institutional, and provider factors. METHODS: This is a retrospective population-based cohort study of hysterectomy for endometrial and cervical cancer in Ontario (2000-2017). Surgical approach, clinicopathologic, sociodemographic, institutional, and provider factors were identified through administrative databases. Fisher's exact, χ2 , Wilcoxon rank sum, logistic regression, and Cox proportional hazards modeling were used to explore factors associated with MIS. RESULTS: A total of 27 652 patients were included. In total, 6199/24 264 (26%) endometrial and 842/3388 (25%) cervical cancer patients received MIS. The proportion of MIS to open surgeries increased from <0.1% in 2000 to over 55% in 2017 (odds ratio [OR] = 1.31, confidence interval [CI] = 1.28-1.34). Low-income quintile, rurality, low hospital volume, nonacademic hospital, nongynecologic oncology surgeon, and earlier year of surgeon graduation were associated with reduced odds of MIS (OR < 1). CONCLUSIONS: The uptake of MIS hysterectomy increased steadily over the time period. Receipt of MIS is dependent upon multiple social determinants, provider variables, and systems factors. These disparities raise concern for health equity in Ontario and have significant implications for health systems planning and resource allocation.

Assessment of oncological safety and utility of hysteroscopy in high grade endometrial cancers: Results from an Israel gynecologic oncology group study.

OBJECTIVE: To compare survival measures of women with Stage I high-grade endometrial cancer who underwent either hysteroscopy or a non-hysteroscopic procedure as a diagnostic procedure. STUDY DESIGN: 298 patients with stage I high grade endometrial cancer who underwent surgery between 2002 and 2014. Patients were divided into two groups: hysteroscopy and non-hysteroscopy (curettage or office endometrial biopsy). Clinical, pathological, and survival measures were compared between the groups. High grade histology included endometroid grade -3, uterine serous papillary carcinoma, clear cell carcinoma, and carcinosarcoma. RESULTS: There were 71 patients in the hysteroscopy group and 227 patients in the non-hysteroscopy group. The median follow-up was 52 months (range 12-120 months). There were no differences between the groups in the 5-year recurrence-free survival (73.9 % vs. 79.7 %; p = 0.65), disease-specific survival (79.3 % vs. 83.6 %; p = 0.87), and overall survival (65.7 % vs. 80.3 %; p = 0.35). CONCLUSION: Hysteroscopic diagnosis in women with early-stage and high-grade endometrial cancer does not adversely affect the survival outcomes.

Surgical margin status in relation to surgical approach in the management of early-stage cervical Cancer: A Canadian cervical Cancer collaborative (4C) study.

OBJECTIVE: Surgical margin status in women undergoing surgery for early-stage cervical cancer is an important prognostic factor. We sought to determine whether close (<3 mm) and positive surgical margins are associated with surgical approach and survival. METHODS: This is a national retrospective cohort study of cervical cancer patients treated with radical hysterectomy. Patients with stage IA1/LVSI-Ib2(FIGO 2018) with lesions up to 4 cm at 11 Canadian institutions from 2007 to 2019 were included. Surgical approach included robotic/laparoscopic (LRH), abdominal (ARH) or combined laparoscopic-assisted vaginal/vaginal (LVRH) radical hysterectomy. Recurrence free survival(RFS) and overall survival (OS) were estimated using Kaplan-Meier analysis. Chi-square and log-rank tests were used to compare groups. RESULTS: 956 patients met inclusion criteria. Surgical margins were as follows: negative (87.0%), positive (0.4%) or close <3 mm (6.8%), missing (5.8%). Most patients had squamous histology (46.9%); 34.6% had adenocarcinomas and 11.3% adenosquamous. Most were stage IB (75.1%) and 24.9% were IA. Mode of surgery included: LRH(51.8%), ARH (39.2%), LVRH (8.9%). Predictive factors for close/positive margins included stage, tumour diameter, vaginal involvement and parametrial extension. Surgical approach was not associated with margin status (p = 0.27). Close/positive margins were associated with a higher risk of death on univariate analysis (HR = non calculable for positive and HR = 1.83 for close margins, p = 0.017), but not significant for OS when adjusted for stage, histology, surgical approach and adjuvant treatment. There were 7 recurrences in patients with close margins (10.3%, p = 0.25). 71.5% with positive/close margins received adjuvant treatment. In addition, MIS was associated with a higher risk of death (OR = 2.39, p = 0.029). CONCLUSION: Surgical approach was not associated to close or positive margins. Close surgical margins were associated with a higher risk of death. MIS was associated with worse survival, suggesting that margin status may not be the driver of worse survival in these cases.

Outcomes associated with treatment to all sites of disease in patients with stage IVB cancer of the cervix.

OBJECTIVE: The use of chemoradiation in patients with stage IVB cancer of the cervix was evaluated to determine if definitive treatment offers benefit. METHODS: A database of 546 patients with cancer of the cervix treated between January 2005 and May 2021 at a tertiary academic medical center was reviewed retrospectively to identify patients with stage IVB disease. Log rank test, regression analysis, and the Kaplan-Meier method were used to identify and compare variables and estimate progression free survival and overall survival. RESULTS: Thirty-three patients with stage IVB cervical cancer were identified. Median age was 53 years (range 28-78). Pathology subtypes were squamous cell (n=22, 67%), adenocarcinoma (n=8, 24%), and clear cell (n=3, 9%). Metastases were classified as lymphatic (n=14, 42%) or hematogenous (n=19, 58%). Following treatment to all sites with chemoradiotherapy and selected use of surgery (n=23), six patients (26%, lymphatic n=4, hematogenous n=2) remained disease free for a median duration of 4 years (range 3-17 years). Recurrences in the remaining patients were distant (n=13) or local (n=4). All patients in the chemotherapy group (n=10, 100%) progressed. Kaplan-Meier analysis showed that median progression free survival was longer for patients treated at all disease sites than for patients treated with chemotherapy alone (19 vs 11 months, p=0.01). However, this was not the case for overall survival (49 vs 33 months, p=0.15). Patients with metastases limited to lymph nodes also had longer median progression free survival (22 vs 11 months, p=0.04) but not overall survival (p=0.68). CONCLUSIONS: Patients with stage IVB cancer of the cervix may benefit from treatment to all sites of disease, if feasible and safe, as demonstrated by improved progression free survival.

Molecular subtype stratified outcomes according to adjuvant therapy in endometrial cancer.

OBJECTIVES: Recent data support the predictive implications of molecular subtype assignment in endometrial cancer (EC). Our objective was to retrospectively assess clinical outcomes according to adjuvant treatment received within EC molecular subtypes. METHODS: Clinical outcomes (disease-specific and progression-free survival DSS/PFS) of EC patients from a single institution and population-based cohorts that had undergone molecular classification were assessed with respect to adjuvant therapy received and 2016 ESMO risk group. RESULTS: 2472 ECs were assessed; 184 (7.4%) POLEmut, 638 (25.8%) MMRd, 1223 (49.5%) NSMP and 427 (17.3%) p53abn. N = 774 (34.6%) of the cohort were ESMO 2016 high risk and 109 (4.8%) were advanced or metastatic. In patients with MMRd EC, assessed across and within stage, there was no observed benefit in DSS or PFS with the addition of chemotherapy +/- radiation compared to radiation alone in ESMO high risk (p = 0.694) or ESMO high, advanced, metastatic risk groups combined (p = 0.852). In patients with p53abn EC, adjuvant chemotherapy given with radiation was associated with significantly longer DSS compared to radiation alone in ESMO high risk (p = 0.007) and ESMO high, advanced and metastatic risk groups combined (p = 0.015), even when restricted to stage I disease (p < 0.001) and when compared in serous vs. non-serous histotypes (p = 0.009). CONCLUSIONS: Adjuvant chemotherapy is associated with more favorable outcomes for patients with p53abn EC, including stage I disease and non-serous histotypes, but does not appear to add benefit within MMRd ECs for any stage of disease, consistent with PORTEC-3 molecular subanalysis. Prospective trials, assessing treatment efficacy within molecular subtype are needed, however these 'real-world' data should be considered when discussing adjuvant treatment with patients.

Oncologic Outcomes of Surgically Treated Cervical Cancer with No Residual Disease on Hysterectomy Specimen: A 4C (Canadian Cervical Cancer Collaborative) Working Group Study.

Minimally invasive surgery for the treatment of macroscopic cervical cancer leads to worse oncologic outcomes than with open surgery. Preoperative conization may mitigate the risk of surgical approach. Our objective was to describe the oncologic outcomes in cases of cervical cancer initially treated with conization, and subsequently found to have no residual cervical cancer after hysterectomy performed via open and minimally invasive approaches. This was a retrospective cohort study of surgically treated cervical cancer at 11 Canadian institutions from 2007 to 2017. Cases initially treated with cervical conization and subsequent hysterectomy, with no residual disease on hysterectomy specimen were included. They were subdivided according to minimally invasive (laparoscopic/robotic (MIS) or laparoscopically assisted vaginal/vaginal hysterectomy (LVH)), or abdominal (AH). Recurrence free survival (RFS) and overall survival (OS) were estimated using Kaplan-Meier analysis. Chi-square and log-rank tests were used to compare between cohorts. Within the total cohort, 238/1696 (14%) had no residual disease on hysterectomy specimen (122 MIS, 103 AH, and 13 VLH). The majority of cases in the cohort were FIGO 2018 stage IB1 (43.7%) and underwent a radical hysterectomy (81.9%). There was no statistical difference between stage, histology, and radical vs simple hysterectomy between the abdominal and minimally invasive groups. There were no significant differences in RFS (5-year: MIS/LVH 97.7%, AH 95.8%, p = 0.23) or OS (5-year: MIS/VLH 98.9%, AH 97.4%, p = 0.10), although event-rates were low. There were only two recurrences. In this large study including only patients with no residual cervical cancer on hysterectomy specimen, no significant differences in survival were seen by surgical approach. This may be due to the small number of events or due to no actual difference between the groups. Further studies are warranted.

Grade and Estrogen Receptor Expression Identify a Subset of No Specific Molecular Profile Endometrial Carcinomas at a Very Low Risk of Disease-Specific Death.

Endometrial carcinoma (EC) can be divided into 4 prognostic molecular subtypes, and no specific molecular profile (NSMP) type is the most commonly occurring type (∼50%). Although described as having an intermediate to favorable prognosis, this subtype encompasses pathologically and molecularly diverse tumors. We aimed to identify factors associated with outcomes within the NSMP ECs that might be used to stratify prognosis and direct treatment. Clinicopathologic, immunohistochemical, and genetic features of a large series of NSMP EC were used to identify parameters that could identify the subset associated with a very favorable outcome (disease-specific death rate <5% at 5 years, termed low-risk NSMP). A total of 1110 NSMP ECs were profiled. In a univariate analysis, stage, grade, lymphovascular invasion, estrogen receptor (ER) and progesterone receptor (PR) expression, L1CAM overexpression, and mutations in PIK3CA were associated with disease-specific survival. Two critical features, grade and ER expression, identified a low-risk NSMP subset (grade 1-2, ER-positive [>1%], 84% of cases), which showed a 5-year disease-specific death rate of 1.6% across all stages and 1.4% within stage I. The remaining cases (high-risk NSMPs, grade 3, and/or ER-negative status) were responsible for most of the disease-specific deaths (disease-specific death rate at 5 years, 22.9%; hazard ratio compared with that of low-risk NSMPs: 16.3; 95% CI, 8.4-31.7). Within NSMP EC, the low-risk and high-risk categories were of prognostic significance independent of the stage on a multivariate analysis. Low-grade and ER-positive NSMP ECs are a homogeneous low-risk group associated with an exceptionally favorable prognosis in which de-escalation and/or endocrine therapy strategies can be applied. Grade 3 and/or ER-negative status identifies a high-risk NSMP subset, including rare high-grade histotypes (eg, clear cell, dedifferentiated, and mesonephric-like), responsible for most NSMP-related deaths. Subclassification of NSMPs allows for the category of low-risk EC molecular subtypes to be dramatically expanded because it now includes both POLEmut and the much more common low-risk NSMP EC.

A double-blinded, randomized trial comparing surgeon-administered transversus abdominis plane block with placebo after midline laparotomy in gynecologic oncology surgery.

BACKGROUND: Surgeon-administered transversus abdominis plane block is a contemporary approach to providing postoperative analgesia, and this approach is performed by transperitoneally administering local anesthetic in the plane between the internal oblique and transversus abdominis muscles to target the sensory nerves of the anterolateral abdominal wall. Although this technique is used in many centers, it has not been studied prospectively in patients undergoing a midline laparotomy. OBJECTIVE: This study aimed to evaluate whether surgeon-administered transversus abdominis plane block reduces postoperative opioid requirements and improves clinical outcomes. STUDY DESIGN: In this double-blind, randomized, placebo-controlled trial, patients with a suspected or proven gynecologic malignancy undergoing surgery through a midline laparotomy at 1 Canadian tertiary academic center were randomized to either the bupivacaine group (surgeon-administered transversus abdominis plane blocks with 40 mL of 0.25% bupivacaine) or the placebo group (surgeon-administered transversus abdominis plane blocks with 40 mL of normal saline solution) before fascial closure. The primary outcome was the total dose of opioids (in morphine milligram equivalents) received in the first 24 hours after surgery. The secondary outcomes included opioid doses between 24 and 48 hours, pain scores, postoperative nausea and vomiting, incidence of clinical ileus, time to flatus, and hospital length of stay. The exclusion criteria included contraindications to study medication, history of chronic opioid use, significant adhesions on the anterior abdominal wall preventing access to the injection site, concurrent nonabdominal surgical procedure, and the planned use of neuraxial anesthesia or analgesia. To detect a 20% decrease in opioid requirements with a 2-sided type 1 error of 5% and power of 80%, a sample size of 36 patients per group was calculated. RESULTS: From October 2020 to November 2021, 38 patients were randomized to the bupivacaine arm, and 41 patients were randomized to the placebo arm. The mean age was 60 years, and the mean body mass index was 29.3. A supraumbilical incision was used in 30 of 79 cases (38.0%), and bowel resection was performed in 10 of 79 cases (12.7%). Patient and surgical characteristics were evenly distributed. The patients in the bupivacaine group required 98.0±59.2 morphine milligram equivalents in the first 24 hours after surgery, whereas the patients in the placebo group required 100.8±44.0 morphine milligram equivalents (P=.85). The mean pain score at 4 hours after surgery was 3.1±2.4 (0-10 scale) in the intervention group vs 3.1±2.0 in the placebo group (P=.93). Clinically significant nausea or vomiting was reported in 1 of 38 patients (2.6%) in the intervention group vs 1 of 41 patients (2.4%) in the placebo group (P=.95). Time to first flatus, rates of clinical ileus, and length of stay were similar between groups. Subgroup analysis of patients with a body mass index of <25 and patients who received an infraumbilical incision showed similarly comparable outcomes. CONCLUSION: Surgeon-administered transversus abdominis plane block with bupivacaine was not found to be superior to the placebo intervention in reducing postoperative opioid requirements or improving other postoperative outcomes for patients undergoing a midline laparotomy. These results differed from previous reports evaluating the ultrasound-guided transversus abdominis plane block approach. Surgeon-administered transversus abdominis plane block should not be considered standard of care in postoperative multimodal analgesia.

Proteomic signature for detection of high-grade ovarian cancer in germline BRCA mutation carriers.

No current screening methods for high-grade ovarian cancer (HGOC) guarantee effective early detection for high-risk women such as germline BRCA mutation carriers. Therefore, the standard-of-care remains risk-reducing salpingo-oophorectomy (RRSO) around age 40. Proximal liquid biopsy is a promising source of biomarkers, but sensitivity has not yet qualified for clinical implementation. We aimed to develop a proteomic assay based on proximal liquid biopsy, as a decision support tool for monitoring high-risk population. Ninety Israeli BRCA1 or BRCA2 mutation carriers were included in the training set (17 HGOC patients and 73 asymptomatic women), (BEDOCA trial; ClinicalTrials.gov Identifier: NCT03150121). The proteome of the microvesicle fraction of the samples was profiled by mass spectrometry and a classifier was developed using logistic regression. An independent cohort of 98 BRCA mutation carriers was used for validation. Safety information was collected for all women who opted for uterine lavage in a clinic setting. We present a 7-protein diagnostic signature, with AUC >0.97 and a negative predictive value (NPV) of 100% for detecting HGOC. The AUC of the biomarker in the independent validation set was >0.94 and the NPV >99%. The sampling procedure was clinically acceptable, with favorable pain scores and safety. We conclude that the acquisition of Müllerian tract proximal liquid biopsies in women at high-risk for HGOC and the application of the BRCA-specific diagnostic assay demonstrates high sensitivity, specificity, technical feasibility and safety. Similar classifier for an average-risk population is warranted.

Upfront surgery in patients with epithelial ovarian cancer and enlarged supradiaphragmatic lymph nodes associated with comparable to neoadjuvant chemotherapy.

BACKGROUND: There is little data regarding the optimal approach to advanced epithelial ovarian cancer (EOC) with isolated extra-peritoneal disease in the cardiophrenic lymph nodes. This study assessed whether the prognosis and surgical outcomes are affected by the treatment approach among these patients. MATERIAL AND METHODS: This retrospective cohort study included patients with advanced EOC, who were treated 2012-2020. Computed tomography scans were reviewed for disease extent and the presence of enlarged supradiaphragmatic nodes (SDLN). Demographic, clinical and oncologic data were recorded. Characteristics and outcomes of patients with and without enlarged SDLN were evaluated, and outcomes of patients with enlarged SDLN who underwent upfront surgery and neoadjuvant chemotherapy were compared. RESULTS: Among 71 women, 47 (66%) had enlarged supradiaphragmatic lymph nodes. Groups had similar baseline characteristics. Among 47 women who had enlarged SDLN. There was no significant difference in progression free survival among patients who had upfront cytoreduction compared to those who received neoadjuvant chemotherapy. Only one asymptomatic chest recurrence was observed. CONCLUSION: Patients with enlarged SDLN have comparable outcomes with either upfront surgery or neoadjuvant chemotherapy. Moreover, the frequency of chest recurrences in patients presenting with enlarged SDLN is exceedingly low.

Inguinofemoral sentinel lymph node biopsy by scar injection in vulvar cancer: an assessment of feasibility and long-term outcomes.

BACKGROUND: Performing inguinofemoral sentinel lymph node biopsy for vulvar cancer following a previous vulvar excision, often referred to as 'scar injection', is debated. OBJECTIVE: To assess the feasibility of sentinel lymph node biopsy following scar injection and the long-term outcomes in patients undergoing this procedure. METHODS: We conducted a retrospective observational cohort study of patients with vulvar cancer. We assessed detection rates and outcomes in patients who underwent sentinel lymph node biopsy by scar injection and compared them with patients who had injection around a visible tumor and with patients who had an inguinofemoral lymphadenectomy following previous vulvar excision. Sentinel node detection rates are described per patient and per groin and are compared using Χ2 analysis. Cox regression analysis was used to assess the association of recurrence and survival with surgical technique and recognized pathological variables. RESULTS: Data were analyzed for 173 groins in 97 patients. At least one sentinel lymph node was detected in 162 (94%) groins examined, and detection rate did not differ whether the groin was assessed following tumor injection (n=122, 94%) or scar injection (n=40, 93%; p=0.85). Patients in the scar-injection group had less frequent lymph node metastases (p<0.02), smaller tumors (p<0.001), and more superficial invasion (p<0.02). Median follow-up was 34.7 months (range 0-108). Scar injection was not independently associated with recurrence or death on multivariable analysis, and depth of invasion was the only independent predictor of disease recurrence (hazards ratio (HR)=1.14, p=0.03). Recurrence and survival were also comparable for patients who had a sentinel lymph node biopsy or inguinofemoral lymphadenectomy following previous vulvar excision (log rank p=0.30; p=0.67). CONCLUSIONS: Sentinel lymph node biopsy by scar injection is feasible and demonstrates similar long-term outcomes in patients having scar or tumor injections, and in patients following previous tumor excision undergoing sentinel lymph node biopsy or lymphadenectomy.

Variability in endometrial carcinoma pathology practice: opportunities for improvement with molecular classification.

We assessed the landscape of diagnostic pathology practice and how molecular classification could potentially impact management of patients with endometrial cancer by collecting patient samples, clinicopathologic data, and patient outcomes from EC patients diagnosed in 2016 at 10 Canadian tertiary cancer centers and 19 community hospitals. ProMisE molecular subtype (POLEmut, MMRd, p53abn, No Specific Molecular Profile (NSMP)) was assigned retrospectively. 1357 patients were fully evaluable including 85 POLEmut (6.3%), 380 MMRd (28.0%), 643 NSMP (47.4%), and 249 p53abn ECs (18.3%). Immunohistochemistry (IHC) for MMR proteins was undertaken at the time of primary diagnosis in 2016 in only 42% of the cohort (570/1357; range 3.5-95.4%/center). p53 IHC had only been performed in 21.1% of the cohort (286/1357; range 10.1-41.9%/center). Thus, based on the retrospective molecular subtype assignment, 54.7% (208/380) of MMRd EC had not been tested with MMR IHC (or MSI) and 48.2% (120/249) of p53abn ECs were not tested with p53 IHC in 2016. Molecular subtype diversity within histotypes was profound; most serous carcinomas were p53abn (91.4%), but only 129/249 (51.8%) p53abn EC were serous. Low-grade (Gr1-2) endometrioid carcinomas were mostly NSMP (589/954, 61.7%) but included all molecular subtypes, including p53abn (19/954, 2.0%). Molecular subtype was significantly associated with clinical outcomes (p < 0.001) even in patients with stage I disease (OS p = 0.006, DSS p < 0.001, PFS p < 0.001). Assessment of national pathologic practice in 2016 shows highly variable use of MMR and p53 IHC and demonstrates significant opportunities to improve and standardize biomarker reporting. Inconsistent, non-reflexive IHC resulted in missed opportunities for Hereditary Cancer Program referral and Lynch Syndrome diagnosis, and missed potential therapeutic implications (e.g., chemotherapy in p53abn EC, immune blockade for MMRd EC). Routine integration of molecular subtyping into practice can improve the consistency of EC pathology assessment and classification.

Disparities in the survival of endometrial cancer patients in a public healthcare system: A population-based cohort study.

OBJECTIVE: Social determinants of health (SDH) have been shown to correlate with adverse cancer outcomes. It is unclear if their impact goes beyond behavioral risk or healthcare access. We aimed to evaluate the association of SDH with endometrial cancer outcomes in a public healthcare system. METHODS: A retrospective cohort study of endometrial cancer patients diagnosed between 2009 and 2017 in Ontario, Canada. Clinical and sociodemographic variables were extracted from administrative databases. Validated multifactorial marginalization scores for domains of material deprivation, residential instability and ethnic concentration were used. Associations between marginalization and survival were evaluated using log-rank testing and Cox proportional hazards regression. RESULTS: 20228 women with endometrial cancer were identified. Fewer patients in marginalized communities presented with early disease (70% vs. 76%, p < 0.001) and received surgery (89% vs. 93%, p < 0.001). Overall survival was shorter among marginalized patients (p < 0.001). On multivariable analysis adjusted for patient and disease factors, overall marginalization (HR = 1.22, 95% CI 1.03-1.08), material deprivation (HR = 1.22, 95% CI 1.10-1.35) and residential instability (HR = 1.32, 95% CI 1.19-1.46) were associated with increased risk of death (p < 0.001). CONCLUSIONS: Socioeconomic marginalization is associated with an increased risk of death in endometrial cancer patients. Targetable events in the cancer care pathway should be identified to improve health equity. FUNDING: This study was supported by a grant (#RD-196) from the Hamilton Health Sciences Juravinski Hospital and Cancer Center Foundation.

Prediction of endometrial cancer recurrence by using a novel machine learning algorithm: An Israeli gynecologic oncology group study.

OBJECTIVES: Endometrial cancer is the most common gynecologic malignancy in developed countries. The overall risk of recurrence is associated with traditional risk factors. METHODS: Machine learning was used to predict recurrence among women who were diagnosed and treated for endometrial cancer between 2002 and 2012 at elven university-affiliated centers. The median follow-up time was 5 years. The following data were retrieved from the medical records and fed into the algorithm: age, chronic metabolic diseases, family and personal cancer history, hormone replacement therapy use, endometrial thickness, uterine polyp presence, complete blood count results, albumin, Ca-125 level, surgical staging, histology, depth of myometrial invasion, LVSI, grade, pelvic washing cytology, and adjuvant treatment. We used XGBoost algorithm, which fits the training data using decision trees, and can also rate the factors according to their influence on the prediction. RESULTS: 1935 women were identified of whom 325 had recurrent disease. On the randomly picked samples, the specificity was 55% and the sensitivity was 98%. Our model showed an operating characteristic curve with AUC of 0.84. CONCLUSIONS: A machine learning algorithm presented promising ability to predict recurrence of endometrial cancer. The algorithm provides an opportunity to identify at-risk patients who may benefit from adjuvant therapy, tighter surveillance, and early intervention.

Predicting recurrence and recurrence-free survival in high-grade endometrial cancer using machine learning.

OBJECTIVE: To develop machine-learning models to predict recurrence and time-to-recurrence in high-grade endometrial cancer (HGEC) following surgery and tailored adjuvant treatment. METHODS: Data were retrospectively collected across eight Canadian centers including 1237 patients. Four models were trained to predict recurrence: random forests, boosted trees, and two neural networks. Receiver operating characteristic curves were used to select the best model based on the highest area under the curve (AUC). For time to recurrence, we compared random forests and Least Absolute Shrinkage and Selection Operator (LASSO) model to Cox proportional hazards. RESULTS: The random forest was the best model to predict recurrence in HGEC; the AUCs were 85.2%, 74.1%, and 71.8% in the training, validation, and test sets, respectively. The top five predictors were: stage, uterus height, specimen weight, adjuvant chemotherapy, and preoperative histology. Performance increased to 77% and 80% when stratified by Stage III and IV, respectively. For time to recurrence, there was no difference between the LASSO and Cox proportional hazards models (c-index 71%). The random forest had a c-index of 60.5%. CONCLUSIONS: A bootstrap random forest model may be a more accurate technique to predict recurrence in HGEC using multiple clinicopathologic factors. For time to recurrence, machine-learning methods performed similarly to the Cox proportional hazards model.

Comparison of outcomes between abdominal, minimally invasive and combined vaginal-laparoscopic hysterectomy in patients with stage IAI/IA2 cervical cancer: 4C (Canadian Cervical Cancer Collaborative) study.

OBJECTIVE: Although minimally invasive hysterectomy (MIS-H) has been associated with worse survival compared to abdominal hysterectomy (AH) for cervical cancer, only 8% of patients in the LACC trial had microinvasive disease (Stage IA1/IA2). We sought to determine differences in outcome among patients undergoing MIS-H, AH or combined vaginal-laparoscopic hysterectomy (CVLH) for microinvasive cervical cancer. METHODS: A retrospective cohort study of all patients undergoing hysterectomy (radical and non radical) for FIGO 2018, microinvasive cervical cancer across 10 Canadian centers between 2007 and 2019 was performed. Recurrence free survival (RFS) was estimated using Kaplan Meier Survival analysis. Chi-square and log-rank tests were used to compare outcomes. RESULTS: 423 patients with microinvasive cervical cancer were included; 259 (61.2%) Stage IA1 (22/8.5% with LVSI) and 164(38.8%) IA2. The median age was 44 years (range 24-81). The most frequent histology was squamous (59.4%). Surgical approach was: 50.1% MIS-H (robotic or laparoscopic), 35.0% AH and 14.9% CVLH. Overall, 70.9% underwent radical hysterectomy and 76.5% had pelvic lymph node assessment. There were 16 recurrences (MIS-H:4, AH:9, CVLH: 3). No significant difference in 5-year RFS was found (96.7% MIS-H, 93.7% AH, 90.0% CVLH, p = 0.34). In a sub-analysis of patients with IA1 LVSI+/IA2(n = 186), survival results were similar. Further, there was no significant difference in peri-operative complications (p = 0.19). Patients undergoing MIS-H had a shorter median length of stay(0 days vs 3 (AH) vs. 1.5 (CVLH), p < 0.001), but had more ER visits (16.0% vs 3.6% (AH), 3.5% (CVLH), p = 0.036). CONCLUSION: In this cohort, including only patients with microinvasive cervical cancer, no difference in recurrence was found by surgical approach. This may be due to the low rate of recurrence making differences hard to detect or due to a true lack of difference. Hence, this patient population may benefit from MIS without compromising oncologic outcomes.

Does sentinel lymph node biopsy in endometrial cancer surgery have an impact on the rate of adjuvant post operative pelvic radiation? An Israeli Gynecologic Oncology Group Study.

Objective: To compare the rates of post-operative radiotherapy between two methods of lymph nodes assessment during surgical staging for endometrial cancer (EC). Methods: We conducted a comparative study of all consecutive women with endometrial cancer who underwent sentinel lymph node detection and biopsy using blue dye and isotope scan (SLNB) at Kaplan Medical Center and patients from the IGOG database, who underwent staging lymphadenectomy (PLND). The primary outcome was the rate of adjuvant and therapeutic radiation. The secondary outcome was a comparison of disease-free survival (DFS) and overall survival (OS). Results: There were 138 patients in the SLNB group and 1022 women in the PLND group. The detection rate of SLN was 74% for unilateral detection and 54% for bilateral detection. In the PLND group 57% were high risk patients vs. 47% in SLNB group (p = 0.03). 43% of high-risk patients in the PLND group received adjuvant or therapeutic pelvic radiation vs. 28% of high-risk women in the SLNB arm (p = 0.017). No statistically significant difference in recurrence rates nor in death rates had been observed in the high-risk group patients. The 5-years survival in the high-risk PLND group was 80% and the recurrence rate was 19% vs. 75% 5-year survival and 14% recurrence in high-risk SLNB cohort, log-rank p = 0.82 for survival and long-rank p = 0.25 for recurrence. Conclusion: Endometrial cancer patients undergoing lymph node assessment by sentinel lymph node biopsy, receive less pelvic radiotherapy.

The impact of preoperative imaging on wait times, surgical approach and overall survival in endometrioid endometrial cancers.

OBJECTIVES: Routine preoperative axial imaging studies (CT/MRI) are not recommended for endometrioid endometrial cancer as they are unlikely to change management and may delay surgery. This study evaluated the association of receiving preoperative imaging on various outcomes. METHODS: A population-based cohort of Endometrioid Endometrial Cancer cases from 2006 to 2016 were identified from the Cancer Registry in Ontario, Canada. Wait time to surgery, type of surgery and overall survival were evaluated in patients with and without preoperative imaging. Predictive factors for wait time > 56 days and aggressive surgery (radical hysterectomy / lymphadenectomy) were determined using multivariable regression analysis. RESULTS: 13,050 cases were included. 22.6% of patients received preoperative imaging, mainly CT scans. Most patients (95.9%) received no neoadjuvant treatment. Patients with preoperative imaging were more likely to have neoadjuvant treatment (11.7% vs. 1.8%) and less likely to have surgery at 180 days post diagnosis (87.9% vs 94.6%). Patients with preoperative imaging had median wait time to surgery of 64 days (47-87), compared to 53 days (36-74) than those without imaging (p < 0.001). Multivariable modeling showed preoperative imaging was associated with decreased odds of having surgery within 56 days (OR = 0.68, 95% CI = 0.62-0.75), and increased odds (OR = 1.73, 95% CI = 1.53-1.95) of having aggressive surgery. The 5-year overall survival for patients with imaging was 84.8% versus 91.1% for patients without preoperative imaging. CONCLUSIONS: Preoperative imaging was associated with longer wait times to surgery, more aggressive surgery, surgery with a gynecologic oncologist and increased use of neoadjuvant and adjuvant treatment. In early-stage disease there was no observed improvement in overall survival for patients with preoperative imaging. Further research on potential benefits of preoperative imaging in higher risk patients is required.