| Background:Artificial Intelligence(AI)has been developing rapidly in the field of medicine,and was gradually moving from laboratory researches to clinical practice.Watson for Oncology(WFO),which was developed by the International Business Machines Corporation(IBM),has been be put into clinical utilityin China since 2017.WFO is an outstanding representative of AI in the medical field,which can provide standardized treatment options for patients quickly and accurately.In recent years,WFO has been applied in more than ten countries,and has been more and more widely used in China as well.However,the number of researches concerning the clinical use of WFO in lung cancer is still limited in China.Therefore,we conducted this retrospective study to examine the concordance between the therapeutic strategies proposed by WFO and the clinical decisions made by oncologists in our cancer center.Objective: The purpose of this study is to explore the feasibility of AI in Chinese lung cancer patients,to analyze the existing disadvantages and advantages,and to provide the basis for future solutions.Methods: We randomly selected lung cancer patients who received anti-tumor treatments in the medical center for precision oncology of The Affiliated Hospital of Medical College Qingdao University from April,2017 to October,2017(N=121).Among all patients,a number of 100 cases met the enrollment criteria according to the WFO standard.The therapeutic options provided by the WFO system could be divided into three categories: recommended,for consideration,and not recommended.Overall,the actual therapeutic regimens proposed by physicians would be defined as concordant with WFO if they corresponded to the recommended or consideration categories and as non-concordant if they corresponded to the not recommended or not available categories Microsoft Excel was used to conduct descriptive statistics analysis patients' baseline characteristics,and we subgrouped all patients enrolled in our study according to histology,gender,age,with/without prior surgery,and analyze the consistency between treatment options recommended by WFO and the actual therapeutic regimens proposed by physicians in each sub group.In addition,we used SPSS 17.0 for statistical analysis.Logistic regression model was used to estimate the row probability ratio and 95% confidence interval of the above factors.P < 0.05 was considered statistically significant.Result: 1.Among all our samples,21 patients didn't meet the inclusion acriteria of WFO.As for the 100 applicable cases,the median age was 61,and 70% of them are male,while 30% are female.Besides,21% of these applicable cases had undergone treatment with surgery,while 79% had not.19% of these 100 patients were diagnosed with small-cell lung cancer and 81% of all had non-small-cell lung cancer.2.Among all the enrolled samples,85% of the treatment options provided by WFO was consistent with those of the actual therapeutic regimens proposed by physicians in our medical center.The clinical consistency between WFO and clinician practice of patients with small cell lung cancer(SCLC)and patients with non-small cell lung cancer(NSCLC)was 89.48% and 83.96% respectively.When it comes to tumor stage,the consistencies in the subgroup of stage II patients,stage III patients and stage IV patients were 83.33%,83.33% and 85.94% respectively.3.When subgrouping the patients based on the gender,the consistency of the diagnosis and treatment strategies proposed by WFO and those provided by phisicians in our medical center was 88.57% in male patients,and 76.67% in female patients.4.When subgrouping the patients based on whether they underwent prior surgery or not,the consistency of the diagnosis and treatment strategies proposed by WFO and those provided by phisicians in our medical center was 85.72% in patients with prior surgery and 84.82% in patients without prior surgery.5.When subgrouping the patients based on age,the diagnosis and treatment strategies proposed by WFO of87.93% of patients aged over and equal to 60 years old were consistent with those provided by phisicians in our medical center,and the diagnosis and treatment strategies proposed by WFO of 80.95% of patients younger than 60 years old were consistent with those provided by phisicians in our medical center.6.The proportions of patients with squamous cell carcinoma and adenocarcinoma are 28.40% and 71.60% respectively.In 86.95% of patients with squamous cell carcinoma and 82.76% of patients with adenocarcinoma,the treatment options suggested by WFO were consistent with those recommended by the physicians in our medical center.Then we further divided patients with adenocarcinoma into the EGFR-mutated group,the EGFR wild-type group,and the EGFR-undetected group.The consistency was 73.34% in the EGFR mutated group,85.71% in the EGFR wild-type group,and 86.66% in the EGFR undetected group.7.Patients with small-cell lung cancer were divided into limited stage and extensive stage.And the diagnosis and treatment strategies proposed by WFO of 77.78% of the patients in limited stage were consistent with those provided by the phisicians in our medical center,while in the subgroup of patients with small cell lung cancer inextensive stage,the concordance rate was as high as100%.8.According to logistic regression model analysis,clinical stages,histological types,gender,age and other factors of lung cancer had no significant effect on the consistency.Conclusions: 1.In China,the treatment strategies proposed by WFO for the enrolled cases were in high consistency with those proposed by the clinical oncologist.2.The consistency of recommendations of AI and options of clinical oncologists in the treatment of lung cancer treatment may be influenced by the constitution of east and west patients,drug availability,medical insurance plan and other factors.3.It is necessary for the AI system to accelerate its localization and improve its consistency with therapeutic decisions made by clinical oncologists,so that it can be more widespreadingly used in clinical or actice in China. |
PDF Full Text Request