Research On The Consistency Of Watson For Oncology In Adjuvant Treatment Of Breast Cancer After Surgery

Objective To study the consistency of Watson for oncology(WFO)and clinicians in Cancer Hosiptal of NingXia medical university for adjuvant treatment of breast cancer after surgery and explore its application value.To analyze the influencing factors of consistency between WFO and clinician treatment decision,and to find reasons of inconsistencies between WFO and clinicians’ treatment decisions.Methods The clinical data of 263 patients who were diagnosed with breast cancer after pathological diagnosis in Cancer Hosiptal of NingXia medical university from January 2017 to January 2018 were collected.WFO provided treatment plans for the identical cases,and the actual treatment plan of clinicians in the Cancer Hosiptal of NingXia medical university were recorded.Using Microsoft Excel descriptive statistical analysis to analyze the consistency,Chi-square test and logistic regression test are used to analyze the influencing factors of consistency.Results 1.Among the 263 adjuvant treatments for breast cancer patients after surgery,there are 211 cases in which WFO and clinicians in Cancer Hosiptal of NingXia medical university have the same actual treatment,the consistency attached 80.2%.The consistency of postoperative adjuvant endocrine therapy was99.3%,the consistency of adjuvant radiotherapy was 96.62%,the consistency of adjuvant targeted therapy was 89.23%,and the consistency of adjuvantchemotherapy was 85.17%.2.As for the 263 breast cancer patients after surgery,WFO and clinicians both suggested chemotherapy at the same time in 224 cases,and 39 patients were inconsistent.The reason is that WFO recommends the Onco type DX when deciding the treatment method,but the clinician adopts the corresponding chemotherapy method according to the characteristics of the clinical case.Further comparing the consistency between WFO and the clinicians’ use of chemotherapy,it was found that the consistency of patients with low-risk breast cancer recurrence risk was lower than those with medium-risk recurrence risk(51.61% VS 95.52%),and the difference was statistically significant(P < 0.001).3.Among the 263 patients with breast cancer after surgery,the consistency of overall treatment programs between the WFO and the clinician was 90.5%,the consistency of the radiotherapy was 100%,the consistency of the endocrine treatment was 100%,and the consistency of the chemotherapy was 95.53%.The consistency of the targeted treatment was 82.75%.There are 58 patients in need of targeted therapy,clinicians all used trastuzumab for targeted therapy,while the WFO recommended trastuzumab for 48 of them,and 10 for trastuzumab and Pertuzumab dual target therapy.4.Univariate analysis found that in adjuvant treatment of breast cancer after surg-ery,menopausal status,molecular typing,Ki-67 level,histological grade,lymph node status,tumor staging were related to the consistency of WFO with the actual treatme-nt plan of clinicians(P<0.05).Logistic multi variate regression model found menopaus-al status(P=0.012,OR:2.89,95%CI:1.26-6.63),histological grade(P=0.019,OR:0.217,95%CI: 0.061-0.781),tumor stag e(P=0.042,OR:0.217,95%CI:0.05-0.943)were independent in-fluencing factors of consistency.Conclusion 1.WFO has well consistency with the clinicians of the Cancer Hosiptal of NingXia medical university in the adjuvant treatment plan for breast cancer after surgery.As for primary hospital,it may become a tool for diagnosis and treatment in the future.3.Menopausal status,molecular typing,lymph node status,histological grade,Ki-67 level,tumor staging are related to the consistency of the actual treatment plan of clinicians in Cancer Hosiptal of NingXia medical university and WFO.Premenopause,higher histological grade,tumor Postoperative staging of breast cancer patients have greater consistency in the use of WFO.3.Patients with low risk of recurrence are significantly different from WFO when choosing whether to receive chemotherapy.The reason is that Onco type DX for breast cancer after surgery are not used to guide the choice of treatment in the clinic.4.The reasons for the actual treatment decisions of clinicians in Cancer Hosiptal of NingXia medical university and WFO are not completely consistent are as follows: Onco type DX after breast cancer surgery,the version of WFO and the availability of drugs,the different guidelines between Eastern and Western,and the patients’ personal wishes and economic situation.In the future,it is necessary to develop artificial intelligence-assisted decision-making systems suitable for China’s national conditions in clinical applications.