PURPOSE Early detection of ovarian cancer, the deadliest gynecologic cancer, is crucial for reducing mortality. Current noninvasive risk assessment measures include protein biomarkers in combination with other clinical factors, which vary in their accuracy. Machine learning can be applied to optimizing the combination of these features, leading to more accurate assessment of malignancy. However, the low prevalence of the disease can make rigorous validation of these tests challenging and can result in unbalanced performance. METHODS MIA3G is a deep feedforward neural network for ovarian cancer risk assessment, using seven protein biomarkers along with age and menopausal status as input features. The algorithm was developed on a heterogenous data set of 1,067 serum specimens from women with adnexal masses (prevalence = 31.8%). It was subsequently validated on a cohort almost twice that size (N = 2,000). RESULTS In the analytical validation data set (prevalence = 4.9%), MIA3G demonstrated a sensitivity of 89.8% and a specificity of 84.02%. The positive predictive value was 22.45%, and the negative predictive value was 99.38%. When stratified by cancer type and stage, MIA3G achieved sensitivities of 94.94% for epithelial ovarian cancer, 76.92% for early-stage cancer, and 98.04% for late-stage cancer. CONCLUSION The balanced performance of MIA3G leads to a high sensitivity and high specificity, a combination that may be clinically useful for providers in evaluating the appropriate management strategy for their patients. Limitations of this work include the largely retrospective nature of the data set and the unequal, albeit random, assignment of histologic subtypes between the training and validation data sets. Future directions may include the addition of new biomarkers or other modalities to strengthen the performance of the algorithm.
BackgroundConservative management of adnexal mass is warranted when there is imaging-based and clinical evidence of benign characteristics. Malignancy risk is, however, a concern due to the mortality rate of ovarian cancer. Malignancy occurs in 10–15% of adnexal masses that go to surgery, whereas the rate of malignancy is much lower in masses clinically characterized as benign or indeterminate. Additional diagnostic tests could assist conservative management of these patients. Here we report the clinical validation of OvaWatch, a multivariate index assay, with real-world evidence of performance that supports conservative management of adnexal masses.MethodsOvaWatch utilizes a previously characterized neural network-based algorithm combining serum biomarkers and clinical covariates and was used to examine malignancy risk in prospective and retrospective samples of patients with an adnexal mass. Retrospective data sets were assembled from previous studies using patients who had adnexal mass and were scheduled for surgery. The prospective study was a multi-center trial of women with adnexal mass as identified on clinical examination and indeterminate or asymptomatic by imaging. The performance to detect ovarian malignancy was evaluated at a previously validated score threshold.ResultsIn retrospective, low prevalence (N = 1,453, 1.5% malignancy rate) data from patients that received an independent physician assessment of benign, OvaWatch has a sensitivity of 81.8% [95% confidence interval (CI) 65.1–92.7] for identifying a histologically confirmed malignancy, and a negative predictive value (NPV) of 99.7%. OvaWatch identified 18/22 malignancies missed by physician assessment. A prospective data set had 501 patients where 106 patients with adnexal mass went for surgery. The prevalence was 2% (10 malignancies). The sensitivity of OvaWatch for malignancy was 40% (95% CI: 16.8–68.7%), and the specificity was 87% (95% CI: 83.7–89.7) when patients were included in the analysis who did not go to surgery and were evaluated as benign. The NPV remained 98.6% (95% CI: 97.0–99.4%). An independent analysis set with a high prevalence (45.8%) the NPV value was 87.8% (95% CI: 95% CI: 75.8–94.3%).ConclusionOvaWatch demonstrated high NPV across diverse data sets and promises utility as an effective diagnostic test supporting management of suspected benign or indeterminate mass to safely decrease or delay unnecessary surgeries.
Aim: This largest-of-its-kind study evaluated the clinical utility of CA125 and OVA1, commonly used as ovarian tumor markers for assessing the risk of malignancy. The research focused on the ability and utility of these tests to reliably predict patients at low risk for ovarian cancer. Clinical utility endpoints were 12-month maintenance of benign mass status, reduction in gynecologic oncologist referral, avoidable surgical intervention and associated cost savings. Materials & methods: This was a multicenter retrospective review of data from electronic medical records and administrative claims databases. Patients receiving a CA125 or OVA1 test between October 2018 and September 2020 were identified and followed for 12 months using site-specific electronic medical records to assess tumor status and utilization outcomes. Propensity score adjustment was used to control for confounding variables. Payer allowed amounts from Merative MarketScan Research Databases were used to estimate 12-month episode-of-care costs per patient, including surgery and other interventions. Results: Among 290 low-risk OVA1 patients, 99.0% remained benign for 12 months compared with 97.2% of 181 low-risk CA125 patients. The OVA1 cohort exhibited 75% lower odds of surgical intervention in the overall sample of patients (Adjusted OR: 0.251, p ≤ 0.0001), and 63% lower odds of gynecologic oncologist utilization among premenopausal women (Adjusted OR: 0.37, p = 0.0390) versus CA125. OVA1 demonstrated significant savings in surgical interventions ($2486, p ≤ 0.0001) and total episode-of-care costs ($2621, p ≤ 0.0001) versus CA125. Conclusion: This study underscores the utility of a reliably predictive multivariate assay for assessing ovarian cancer risk. For patients assessed at low risk of ovarian tumor malignancy, OVA1 is associated with a significant reduction in avoidable surgeries and substantial cost savings per patient. OVA1 is also associated with a significant reduction in subspecialty referrals for low-risk premenopausal patients.
e17607 Background: Ovarian cancer (OC) – most lethal gynecologic malignancy - is rare among women with adnexal mass. Preventive adnexal surgery is a common practice, despite pathology confirming malignancy in only 10-15% of these cases. This motivates a surveillance-based clinical management paradigm to prognosticate the risk for malignancy and channel appropriately for surgery. We recently developed a multivariate index assay -MIA3G[1] - to assess risk for OC in women with adnexal mass. In a multi-center prospective surveillance trial, we assessed the concordance between preventive surgery and clinical management using MIA3G. Methods: MIA3G risk prediction uses a DNN algorithm which models 7 serum biomarkers (CA125, HE4, BM2, APO, FSH, TFR, PreAlbumin), age and menopausal status. With NPV of 99.7% (CI:99.2–99.9) it risk stratifies patients into low probability of malignancy or indeterminate. Across 11 centers, the MIA3G cohort comprised of 546 women (IQR:41-61yo) with symptomatic or asymptomatic masses and/or presence of HBOC variants. Serial blood draws were timed with clinic-visits and the ultimate one coincident with surgery. The objectives were: 1) prognosticate the MIA3G based preventive surgery pool and 2) establish concordance between the MIA3G probabilistic risk and the pathology-based outcome of malignancy versus benign. Results: In the study cohort of 546 patients, 20.3% (111) underwent preventive surgery. Conservatively, at a 12.8% FPR[1], MIA3G would have stratified approximately 70 patients, a projection of 37% less for surgery. Partitioning by clinical characteristic: (i) symptomatic: In comparison to 27.8% (73/263) of patients referred to surgery, MIA3G projections would have resulted in a 54.8%(33/263) reduction. (ii) asymptomatic: 22.5%(31/138) of patients were referred to surgery; MIA3G projections would have resulted in a 45.2% (17/138) reduction. In the 111 preventive surgeries, pathology confirmed a malignancy rate of 6.3% (7/111). The overall concordance of MIA3G with pathology was 84.68% (94/111) which increased to 91.98% (102/111) in the pre-menopausal (PRE) sub-cohort. The malignant discordancy (MIA3G:FN), all in PRE, comprised of early-stage epithelial: endometreoid, Leiomysarcoma and Sertoli-Leydig cell where the masses were characterized as cystic benign/indeterminate by ultrasound. These findings underscore the power of MIA3G when preventive surgery is considered for patients with ovarian cancer risk. Conclusions: The multivariate index assay MIA3G reduces numbers of preventative surgery for ovarian cancer risk. The high concordance establishes that MIA3G non-invasive surveillance-based clinical management can be effective for the risk assessment of ovarian cancer in patients presented with adnexal mass. Surgical management should be reserved for high probability of malignancy. [1]: DOI 10.3389/fmed.2023.1102437.
e17608 Background: Monitoring of ovarian malignancy risk in indeterminate masses is typically done using TVUS and other clinical factors. No biomarkers are available to assist physicians in clinical management of benign or indeterminate adnexal mass either on initial visit or in follow-up. We recently characterized the clinical performance of a deep neural network-derived multivariate index assay (MIA3G) [1] to assess ovarian malignancy risk in retrospective and prospectively collected multi-site clinical studies. Here we present data on MIA3G serial follow-up to monitor ovarian cancer risk in women presented with adnexal mass from the prospective studies. Methods: Data are presented from ongoing multisite clinical studies in which total 924 patients presented with adnexal masses were enrolled. Follow-up visits, scheduled at clinician’s discretion, may have included imaging and blood collection. Specimens were processed for serum and run on a clinical analyzer. MIA3G score from 0-10 was calculated using seven serum biomarkers coupling with patient age and menopausal status based on previously published neural network-based algorithm. MIA3G with NPV of 99.7% (CI:99.2–99.9) was used to risk stratify the patient with an adnexal mass into low probability of malignancy or indeterminate with a validated cut off at 5.0. For this analysis, MIA3G scores were binned into the following Zones: I (0-2.49), II (2.50 -4.99) and II (5.00-10.00). Results: Of 924 enrolled patients, 538 patients had completed clinical and biomarker data on initial study draw. Of these, 145 had at least one follow-up test and 31 had at least two follow-up test. Median duration to first follow-up test was 108 d (~3.6 mon), and median duration to second follow-up test was 272 d (~9.1 mon). Follow-up in 3 MIA3G score zones at the 0 -3.6 and 3.6 - 9.1 mon interval consistently showed 88% patients within zone I remained unchanged, whereas 12% of patients moved to zones II and III. In the 0 - 3.6 mon interval, the median change in MIA3G score from I to II was 2.25 (n = 9, range 0.54 - 3.52) and from I to III was 5.92 (n = 8, range 5.08-7.62 ). In the 3.6 - 9.1 mon testing interval, the median change in MIA3G score from zone I to II was 3.11 (n = 2, range 2.75 to 3.46) and the single patient change from I to III was 6.89. Across both testing intervals, 17%-50% of patients in zone II remain unchanged while approximately 25% patients had score increase to zone III. Importantly, 50% of patients in zone III remained unchanged in the absence of clinical intervention whereas 50% had MIA3G score reduction in association with clinical management. Conclusions: These results indicate that i) the serial follow-up with MIA3G is recommended to monitor the clinical status of the adnexal mass every 3 months, ii) MIA3G score changes of > 2.25 suggest clinical follow-up, and finally iii) the MIA3G is a suitable tool for the effectiveness of clinical management of adnexal mass. [1]: DOI 10.3389/fmed.2023.1102437.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.