A chemotherapy response classifier based on support vector machines for high-grade serous ovarian carcinoma

Predictive Value of Machine Learning for Platinum Chemotherapy Responses in Ovarian Cancer: Systematic Review and Meta-Analysis

BACKGROUND

Machine learning is a potentially effective method for predicting the response to platinum-based treatment for ovarian cancer. However, the predictive performance of various machine learning methods and variables is still a matter of controversy and debate.

OBJECTIVE

This study aims to systematically review relevant literature on the predictive value of machine learning for platinum-based chemotherapy responses in patients with ovarian cancer.

METHODS

Following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, we systematically searched the PubMed, Embase, Web of Science, and Cochrane databases for relevant studies on predictive models for platinum-based therapies for the treatment of ovarian cancer published before April 26, 2023. The Prediction Model Risk of Bias Assessment tool was used to evaluate the risk of bias in the included articles. Concordance index (C-index), sensitivity, and specificity were used to evaluate the performance of the prediction models to investigate the predictive value of machine learning for platinum chemotherapy responses in patients with ovarian cancer.

RESULTS

A total of 1749 articles were examined, and 19 of them involving 39 models were eligible for this study. The most commonly used modeling methods were logistic regression (16/39, 41%), Extreme Gradient Boosting (4/39, 10%), and support vector machine (4/39, 10%). The training cohort reported C-index in 39 predictive models, with a pooled value of 0.806; the validation cohort reported C-index in 12 predictive models, with a pooled value of 0.831. Support vector machine performed well in both the training and validation cohorts, with a C-index of 0.942 and 0.879, respectively. The pooled sensitivity was 0.890, and the pooled specificity was 0.790 in the training cohort.

CONCLUSIONS

Machine learning can effectively predict how patients with ovarian cancer respond to platinum-based chemotherapy and may provide a reference for the development or updating of subsequent scoring systems.

Gynecological cancer prognosis using machine learning techniques: A systematic review of last three decades (1990–2022)

OBJECTIVE

Many Computer Aided Prognostic (CAP) systems based on machine learning techniques have been proposed in the field of oncology. The objective of this systematic review was to assess and critically appraise the methodologies and approaches used in predicting the prognosis of gynecological cancers using CAPs.

METHODS

Electronic databases were used to systematically search for studies utilizing machine learning methods in gynecological cancers. Study risk of bias (ROB) and applicability were assessed using the PROBAST tool. 139 studies met the inclusion criteria, of which 71 predicted outcomes for ovarian cancer patients, 41 predicted outcomes for cervical cancer patients, 28 predicted outcomes for uterine cancer patients, and 2 predicted outcomes for gynecological malignancies broadly.

RESULTS

Random forest (22.30 %) and support vector machine (21.58 %) classifiers were used most commonly. Use of clinicopathological, genomic and radiomic data as predictors was observed in 48.20 %, 51.08 % and 17.27 % of studies, respectively, with some studies using multiple modalities. 21.58 % of studies were externally validated. Twenty-three individual studies compared ML and non-ML methods. Study quality was highly variable and methodologies, statistical reporting and outcome measures were inconsistent, preventing generalized commentary or meta-analysis of performance outcomes.

CONCLUSION

There is significant variability in model development when prognosticating gynecological malignancies with respect to variable selection, machine learning (ML) methods and endpoint selection. This heterogeneity prevents meta-analysis and conclusions regarding the superiority of ML methods. Furthermore, PROBAST-mediated ROB and applicability analysis demonstrates concern for the translatability of existing models. This review identifies ways that this can be improved upon in future works to develop robust, clinically translatable models within this promising field.

A Subset of Secreted Proteins in Ascites Can Predict Platinum-Free Interval in Ovarian Cancer

The time between the last cycle of chemotherapy and recurrence, the platinum-free interval (PFI), predicts overall survival in high-grade serous ovarian cancer (HGSOC). To identify secreted proteins associated with a shorter PFI, we utilized machine learning to predict the PFI from ascites composition. Ascites from stage III/IV HGSOC patients treated with neoadjuvant chemotherapy (NACT) or primary debulking surgery (PDS) were screened for secreted proteins and Lasso regression models were built to predict the PFI. Through regularization techniques, the number of analytes used in each model was reduced; to minimize overfitting, we utilized an analysis of model robustness. This resulted in models with 26 analytes and a root-mean-square error (RMSE) of 19 days for the NACT cohort and 16 analytes and an RMSE of 7 days for the PDS cohort. High concentrations of MMP-2 and EMMPRIN correlated with a shorter PFI in the NACT patients, whereas high concentrations of uPA Urokinase and MMP-3 correlated with a shorter PFI in PDS patients. Our results suggest that the analysis of ascites may be useful for outcome prediction and identified factors in the tumor microenvironment that may lead to worse outcomes. Our approach to tuning for model stability, rather than only model accuracy, may be applicable to other biomarker discovery tasks.

Applying Artificial Intelligence to Gynecologic Oncology: A Review

Importance

Artificial intelligence (AI) will play an increasing role in health care. In gynecologic oncology, it can advance tailored screening, precision surgery, and personalized targeted therapies.

Objective

The aim of this study was to review the role of AI in gynecologic oncology.

Evidence Acquisition

Artificial intelligence publications in gynecologic oncology were identified by searching "gynecologic oncology AND artificial intelligence" in the PubMed database. A review of the literature was performed on the history of AI, its fundamentals, and current applications as related to diagnosis and treatment of cervical, uterine, and ovarian cancers.

Results

A PubMed literature search since the year 2000 showed a significant increase in oncology publications related to AI and oncology. Early studies focused on using AI to interrogate electronic health records in order to improve clinical outcome and facilitate clinical research. In cervical cancer, AI algorithms can enhance image analysis of cytology and visual inspection with acetic acid or colposcopy. In uterine cancers, AI can improve the diagnostic accuracies of radiologic imaging and predictive/prognostic capabilities of clinicopathologic characteristics. Artificial intelligence has also been used to better detect early-stage ovarian cancer and predict surgical outcomes and treatment response.

Conclusions and Relevance

Artificial intelligence has been shown to enhance diagnosis, refine clinical decision making, and advance personalized therapies in gynecologic cancers. The rapid adoption of AI in gynecologic oncology will depend on overcoming the challenges related to data transparency, quality, and interpretation. Artificial intelligence is rapidly transforming health care. However, many physicians are unaware that this technology is being used in their practices and could benefit from a better understanding of the statistics and computer science behind these algorithms. This review provides a summary of AI, its applicability, and its limitations in gynecologic oncology.

BRCA1 Expression by Immunohistochemistry and Prognosis in Ovarian Cancer: A Systematic Review and Meta-Analysis

BACKGROUND

Homologous recombination deficiencies are associated with increased platinum sensitivity and potential response to poly (ADP-ribose) polymerase inhibitors in epithelial ovarian cancer. As an alternative to germline testing or somatic tumor sequencing, BRCA1 deficiency can be detected by immunohistochemistry and might predict homologous recombination deficiencies.

OBJECTIVE

This study aimed to assess the association between BRCA1 expression by immunohistochemistry and the prognosis of patients with epithelial ovarian cancer.

METHODS

We conducted a systematic review and meta-analysis following the Preferred Reporting Items for Systematic reviews and Meta-Analyses statement. We searched PubMed, EMBASE, Web of Science, and Scopus databases through July 2019. Reference lists of selected articles were screened for further studies. We conducted qualitative synthesis and meta-analyses of hazard ratios for overall survival and progression-free survival.

RESULTS

Of 41 studies of BRCA1 expression using immunohistochemistry, 18 evaluated the association of BRCA1 expression with patient survival (2738 cases). The loss of BRCA1 expression was associated with improved overall survival (hazard ratio = 0.67, 95% confidence interval 0.57-0.77) and progression-free survival (hazard ratio = 0.70, 95% confidence interval 0.58-0.84).

CONCLUSIONS

Negative BRCA1 expression assessed by immunohistochemistry was associated with a better prognosis in epithelial ovarian cancer.

Artificial Intelligence in Pharmacoepidemiology: A Systematic Review. Part 1-Overview of Knowledge Discovery Techniques in Artificial Intelligence

Aim

To perform a systematic review on the application of artificial intelligence (AI) based knowledge discovery techniques in pharmacoepidemiology.

Study Eligibility Criteria

Clinical trials, meta-analyses, narrative/systematic review, and observational studies using (or mentioning articles using) artificial intelligence techniques were eligible. Articles without a full text available in the English language were excluded.

Data Sources

Articles recorded from 1950/01/01 to 2019/05/06 in Ovid MEDLINE were screened.

Participants

Studies including humans (real or simulated) exposed to a drug.

Results

In total, 72 original articles and 5 reviews were identified via Ovid MEDLINE. Twenty different knowledge discovery methods were identified, mainly from the area of machine learning (66/72; 91.7%). Classification/regression (44/72; 61.1%), classification/regression + model optimization (13/72; 18.0%), and classification/regression + features selection (12/72; 16.7%) were the three most frequent tasks in reviewed literature that machine learning methods has been applied to solve. The top three used techniques were artificial neural networks, random forest, and support vector machines models.

Conclusions

The use of knowledge discovery techniques of artificial intelligence techniques has increased exponentially over the years covering numerous sub-topics of pharmacoepidemiology.

Systematic Review Registration

Systematic review registration number in PROSPERO: CRD42019136552.

Immunohistochemistry for the detection of BRCA1 and BRCA2 proteins in patients with ovarian cancer: a systematic review

BACKGROUND

Loss of function in either breast cancer type 1 susceptibility protein (BRCA1) or breast cancer type 2 susceptibility protein (BRCA2) is a major risk factor for epithelial ovarian cancer (EOC) development. BRCA1 or BRCA2 deficiencies are associated with short-term prognosis and might have importance for the treatment of women with the disease. However, the screening of all possible mechanisms of dysfunction is expensive, time-consuming and difficult to apply in clinical practice. On the other hand, immunohistochemistry (IHC) is a simple and reliable method to access the expression of several proteins in tumour tissues.

MATERIALS AND METHODS

This systematic review aims to evaluate the current usage of IHC to detect BRCA1 and BRCA2 deficiencies in EOC. We searched and evaluated all primary literature on the use of IHC for evaluating BRCA1 and BRCA2 proteins expression in EOC. The main concepts for the search were: ovarian neoplasms, IHC, BRCA1 and BRCA2.

RESULTS

Forty-four studies from 925 unique titles were included. A total of 4206 tumour samples were evaluated for BRCA1 and 1041 for BRCA2 expression. Twelve BRCA1 primary antibodies were used in 41 studies, and the most common was the MS110 clone (75.6%). Seven BRCA2 primary antibodies were used in ten studies. Using the cut-off of 10%, 47.0% of EOCs are associated with loss of BRCA1 and 34.5% with the loss of BRCA2 expression.

CONCLUSION

IHC was effective to detect loss of BRCA1 protein expression in EOC; however, data on BRCA2 expression were heterogeneous and difficult to interpret.

A reactive oxygen species scoring system predicts cisplatin sensitivity and prognosis in ovarian cancer patients

Background

To reveal roles of reactive oxygen species (ROS) status in chemotherapy resistance and to develop a ROS scoring system for prognosis prediction in ovarian cancer.

Methods

We tested the sensitizing effects of ROS elevating drugs to cisplatin (cDDP) in ovarian cancer both in vitro and in vivo. A ROS scoring system was developed using The Cancer Genome Atlas (TCGA) database of ovarian cancer. The associations between ROS scores and overall survival (OS) were analyzed in TCGA, Tothill dataset, and our in-house dataset (TJ dataset).

Results

ROS-inducing drugs increased cisplatin-induced ovarian cancer cell injury in vitro and in vivo. ROS scoring system was established using 25 ROS-related genes. Patients were divided into low (scores 0–12) and high (scores 13–25) score groups. Improved patient survival was associated with higher scores (TCGA dataset hazard ratio (HR) = 0.43, P < 0.001; Tothill dataset HR = 0.65, P = 0.022; TJ dataset HR = 0.40, P = 0.003). The score was also significantly associated with OS in multiple datasets (TCGA dataset r² = 0.574, P = 0.032; Thothill dataset r² = 0.266, P = 0.049; TJ dataset r² = 0.632, P = 0.001) and with cisplatin sensitivity in ovarian cancer cell lines (r² = 0.799, P = 0.016) when used as a continuous variable. The scoring system showed better prognostic performance than other clinical factors by receiver operating characteristic (ROC) curves (TCGA dataset area under the curve (AUC) = 0.71 v.s. 0.65, Tothill dataset AUC = 0.73 v.s. 0.67, TJ dataset AUC = 0.74 v.s. 0.66).

Conclusions

ROS status is associated with chemotherapy resistance. ROS score system might be a prognostic biomarker in predicting the survival benefit from ovarian cancer patients.