Predicting complete cytoreduction for advanced ovarian cancer patients using nearest-neighbor models
Abstract Background The foundation of modern ovarian cancer care is cytoreductive surgery to remove all macroscopic disease (R0). Identification of R0 resection patients may help individualise treatment. Machine learning and AI have been shown to be effective systems for classification and predictio...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2020-09-01
|
| Series: | Journal of Ovarian Research |
| Subjects: | |
| Online Access: | http://link.springer.com/article/10.1186/s13048-020-00700-0 |
| _version_ | 1828075841793294336 |
|---|---|
| author | Alexandros Laios Alexandros Gryparis Diederick DeJong Richard Hutson Georgios Theophilou Chris Leach |
| author_facet | Alexandros Laios Alexandros Gryparis Diederick DeJong Richard Hutson Georgios Theophilou Chris Leach |
| author_sort | Alexandros Laios |
| collection | DOAJ |
| description | Abstract Background The foundation of modern ovarian cancer care is cytoreductive surgery to remove all macroscopic disease (R0). Identification of R0 resection patients may help individualise treatment. Machine learning and AI have been shown to be effective systems for classification and prediction. For a disease as heterogenous as ovarian cancer, they could potentially outperform conventional predictive algorithms for routine clinical use. We investigated the performance of an AI system, the k-nearest neighbor (k-NN) classifier, to predict R0, comparing it with logistic regression. Patients diagnosed with advanced stage, high grade serous ovarian, tubal and primary peritoneal cancer, undergoing surgical cytoreduction from 2015 to 2019, was selected from the ovarian database. Performance variables included age, BMI, Charlson Comorbidity Index, timing of surgery, surgical complexity and disease score. The k-NN algorithm classified R0 vs non-R0 patients using 3–20 nearest neighbors. Prediction accuracy was estimated as percentage of observations in the training set correctly classified. Results 154 patients were identified, with mean age of 64.4 + 10.5 yrs., BMI of 27.2 + 5.8 and mean SCS of 3 + 1 (1–8). Complete and optimal cytoreduction was achieved in 62 and 88% patients. The mean predictive accuracy was 66%. R0 resection prediction of true negatives was as high as 90% using k = 20 neighbors. Conclusions The k-NN algorithm is a promising and versatile tool for R0 resection prediction. It slightly outperforms logistic regression and is expected to improve accuracy with data expansion. |
| first_indexed | 2024-04-11T02:05:34Z |
| format | Article |
| id | doaj.art-cd5afab14dda4c77ba58f7483849be2c |
| institution | Directory Open Access Journal |
| issn | 1757-2215 |
| language | English |
| last_indexed | 2024-04-11T02:05:34Z |
| publishDate | 2020-09-01 |
| publisher | BMC |
| record_format | Article |
| series | Journal of Ovarian Research |
| spelling | doaj.art-cd5afab14dda4c77ba58f7483849be2c2023-01-03T03:19:09ZengBMCJournal of Ovarian Research1757-22152020-09-011311810.1186/s13048-020-00700-0Predicting complete cytoreduction for advanced ovarian cancer patients using nearest-neighbor modelsAlexandros Laios0Alexandros Gryparis1Diederick DeJong2Richard Hutson3Georgios Theophilou4Chris Leach5Department of Gynaecological Oncology, St James’s University Hospital, Leeds Teaching HospitalsUnit of Endocrinology, Diabetes Mellitus and Metabolism, Aretaion Hospital, National and Kapodistrian University of Athens School of MedicineDepartment of Gynaecological Oncology, St James’s University Hospital, Leeds Teaching HospitalsDepartment of Gynaecological Oncology, St James’s University Hospital, Leeds Teaching HospitalsDepartment of Gynaecological Oncology, St James’s University Hospital, Leeds Teaching HospitalsSchool of Human & Health Sciences, University of HuddersfieldAbstract Background The foundation of modern ovarian cancer care is cytoreductive surgery to remove all macroscopic disease (R0). Identification of R0 resection patients may help individualise treatment. Machine learning and AI have been shown to be effective systems for classification and prediction. For a disease as heterogenous as ovarian cancer, they could potentially outperform conventional predictive algorithms for routine clinical use. We investigated the performance of an AI system, the k-nearest neighbor (k-NN) classifier, to predict R0, comparing it with logistic regression. Patients diagnosed with advanced stage, high grade serous ovarian, tubal and primary peritoneal cancer, undergoing surgical cytoreduction from 2015 to 2019, was selected from the ovarian database. Performance variables included age, BMI, Charlson Comorbidity Index, timing of surgery, surgical complexity and disease score. The k-NN algorithm classified R0 vs non-R0 patients using 3–20 nearest neighbors. Prediction accuracy was estimated as percentage of observations in the training set correctly classified. Results 154 patients were identified, with mean age of 64.4 + 10.5 yrs., BMI of 27.2 + 5.8 and mean SCS of 3 + 1 (1–8). Complete and optimal cytoreduction was achieved in 62 and 88% patients. The mean predictive accuracy was 66%. R0 resection prediction of true negatives was as high as 90% using k = 20 neighbors. Conclusions The k-NN algorithm is a promising and versatile tool for R0 resection prediction. It slightly outperforms logistic regression and is expected to improve accuracy with data expansion.http://link.springer.com/article/10.1186/s13048-020-00700-0Ovarian CancerCytoreductionPredictive factorsMachine learningArtificial intelligence |
| spellingShingle | Alexandros Laios Alexandros Gryparis Diederick DeJong Richard Hutson Georgios Theophilou Chris Leach Predicting complete cytoreduction for advanced ovarian cancer patients using nearest-neighbor models Journal of Ovarian Research Ovarian Cancer Cytoreduction Predictive factors Machine learning Artificial intelligence |
| title | Predicting complete cytoreduction for advanced ovarian cancer patients using nearest-neighbor models |
| title_full | Predicting complete cytoreduction for advanced ovarian cancer patients using nearest-neighbor models |
| title_fullStr | Predicting complete cytoreduction for advanced ovarian cancer patients using nearest-neighbor models |
| title_full_unstemmed | Predicting complete cytoreduction for advanced ovarian cancer patients using nearest-neighbor models |
| title_short | Predicting complete cytoreduction for advanced ovarian cancer patients using nearest-neighbor models |
| title_sort | predicting complete cytoreduction for advanced ovarian cancer patients using nearest neighbor models |
| topic | Ovarian Cancer Cytoreduction Predictive factors Machine learning Artificial intelligence |
| url | http://link.springer.com/article/10.1186/s13048-020-00700-0 |
| work_keys_str_mv | AT alexandroslaios predictingcompletecytoreductionforadvancedovariancancerpatientsusingnearestneighbormodels AT alexandrosgryparis predictingcompletecytoreductionforadvancedovariancancerpatientsusingnearestneighbormodels AT diederickdejong predictingcompletecytoreductionforadvancedovariancancerpatientsusingnearestneighbormodels AT richardhutson predictingcompletecytoreductionforadvancedovariancancerpatientsusingnearestneighbormodels AT georgiostheophilou predictingcompletecytoreductionforadvancedovariancancerpatientsusingnearestneighbormodels AT chrisleach predictingcompletecytoreductionforadvancedovariancancerpatientsusingnearestneighbormodels |