Consistency Of Watson For Oncology Ai-Assisted Decision-Making System In First-Line Treatment Of Advanced Breast Cancer

Objective To explore the consistency of clinical decision-making in the first-line treatment of advanced breast cancer with the recommended methods and schemes of WFO artificial intelligence-assisted diagnosis and treatment system.Analyze the factors that affect its consistency,explore the causes of "inconsistent" diagnosis and treatment decisions,and evaluate the advantages and disadvantages of the two,providing a reference for the future application of the WFO-assisted diagnosis and treatment decision system in clinical work in the region.Methods Collected patients who received first-line treatment of advanced breast cancer from January 2014 to December 2018 at the Ningxia Medical University Cancer Hospital,entered medical history information into the WFO,and obtained a consistent rate.The factors of decision consistency were analyzed,and the non-parametric test Wilcoxon signed rank sum test was used to compare the difference between progression-free survival(PFS)between the WFO recommendation and the real-world treatment.Results 1.Among 200 patients with advanced breast cancer,the consistent rate of WFO and real-world treatment was 50.5%(101 cases),of which 87 patients received chemotherapy,14 patients received endocrine therapy,and a total of 40 patients received targeted therapy combined with endocrine therapy.Or chemotherapy.2.The overall agreement rate of WFO and real-world treatment regimens in patients with advanced breast cancer is 23.5%(47/200).The agreement rate of endocrine therapy regimens(57.4%)is higher than that of chemotherapy(22.3%)and targeted therapy regimens.Rate(11.3%),the difference was statistically significant(X2 = 32.673,P <0.05).3.In the first-line treatment of advanced breast cancer,the agreement rates of WFO and real-world treatment for patients with hormone receptor negative and hormone receptor positive advanced breast cancer were 74% and 42.6%,respectively,and the difference was statistically significant(X2 = 14.728,P <0.05).Age,menstrual status,metastatic site,pathological type,and HER-2 receptor expression had no significant effect on the consistency of WFO with the actual treatment(P> 0.05).4.The rate of agreement between WFO and HER-2 positive and HER-2 negative advanced breast cancer patients in the real world is 1.4% and 18.2%,respectively,and the difference was statistically significant(X2 = 8.004,P <0.05).Further analysis of the main reasons for inconsistency: WFO recommends trastuzumab combined with pertuzumab dual-targeted therapy for HER-2 positive patients,while 23 patients in the real world are treated with trastuzumab monotherapy For economic reasons,20 patients did not use targeted therapy.Age,menstrual status,metastatic site,pathological type,and HR receptor status had no significant effect on the consistency of the treatment plan(P> 0.05).5.By the end of the follow-up(February 1,2020),200 cases of advanced breast cancer patients had 145 cases of recurrence and metastasis in the real world.Among them,the PFS in Luminal patients in the real world was shorter than that shown by the clinical evidence of the WFO recommended protocol(10 months v.s18.7 months),and the difference was statistically significant(Z =-4.566,P <0.05).Further analysis of the results showed that WFO recommended endocrine combined with CDK4 / 6 inhibitor therapy and 61 cases of breast cancer patients who chose chemotherapy sequential endocrine therapy in the real world.There was no statistically significant difference in PFS between HER-2 overexpression and triple negative patients in the real world and clinical evidence from the WFO recommendedprotocol(P> 0.05).Conclusion 1.The agreement between WFO and real-world treatment methods and treatment plans for patients with advanced first-line breast cancer is low.The agreement rate of first-line endocrine treatment is significantly higher than that of chemotherapy and targeted treatment.2.Patients with HR-negative advanced breast cancer are more consistent with real-world treatment than patients with HR-positive.Factors such as age,menstrual status,metastatic site,pathological type,and HER-2 expression did not significantly affect the consistency of WFO with real-world treatment.3.The agreement rate of WFO and real-world treatment in patients with HER-2 positive advanced breast cancer is lower than that in patients with HER-2 negative.Age,menstrual status,metastatic site,pathological type,HR receptor status and other factors have consistent agreement with treatment plan No significant impact.4.Luminal-type advanced breast cancer patients in the real world have a shorter PFS than the PFS suggested by clinical evidence from the WFO recommended protocol.The real-world PFS of HER-2 overexpressed and triple negative breast cancer patients has no significant difference from the clinical evidence of the recommended protocol of WFO.5.The reasons for inconsistent WFO and real-world diagnosis and treatment decisions are:differences in Eastern and Western guidelines,drug accessibility,indications for drug instructions,and physical symptoms,willingness,and economic strength of patients.