The prognostic value of artificial intelligence-based P53 and Ki67 detection in stage Ⅰ non-mucinous adenocarcinoma of lung
10.3760/cma.j.cn101721-20250218-00075
- VernacularTitle:人工智能辅助P53及Ki67计数在肺Ⅰ期INMA患者预后评估中的价值
- Author:
Jiting DI
1
;
Kang QI
;
Dong LI
;
Gang LIN
;
Yan XIONG
Author Information
1. 北京大学第一医院病理科，北京 100034
- Publication Type:Journal Article
- Keywords:
Pulmonary non-mucinous adenocarcinoma;
Artificial intelligence;
P53;
Ki67;
Prognosis
- From:
Clinical Medicine of China
2025;41(6):407-416
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the prognostic value of artificial intelligence-based P53 and Ki67 detection in stage I non-mucinous adenocarcinoma(INMA)of lung.Methods:A retrospective analysis was made of patients treated by radical surgical resection for INMA of lung in the Department of Thoracic Surgery of Peking University First Hospital from Jan.2015 to Dec.2016,with complete clinicopathological and 5-year follow-up data.Immunohistochemical staining for P53 and Ki67 was performed on all cases and the index of P53 and Ki67 was calculated with the assistance of artificial intelligence(AI).The optimal cut-off values for P53 and Ki67 were determined using X-Tile software,and based on these values,the patients were divided into low-expression and high-expression groups.Pearson chi-square test and Fisher's exact test were used to compare the differences in clinicopathological characteristics between the different groups.Univariate and multivariate Cox regression analyses were performed to assess the impact of various indicators on 5-year overall survival(OS)and disease-free survival (DFS)for stage I INMA.The time-dependent receiver operating characteristic(ROC) curves and the area under the curve(AUC)was used to analyze the predictive performance of P53 and Ki67 for the prognosis of stage I INMA.Results:Among the 191 patients, the median follow-up time was 60(54, 60) months. The index of P53 and Ki67 were 0%-100% and 1.0%-78.0%,respectively. The X-Tile software revealed optimal cut-off values of 62% for P53 and 20% for Ki67.Then the patients were divided into P53 low-expression group (<62%), P53 high-expression (≥62%) group and Ki67 low-expression (<20%)group,Ki67 high-expression group (≥20%). High expression of P53 was associated with male ( χ2=12.45, P<0.001), smoking ( χ2=12.24, P<0.001), pTNM stage ( χ2=16.28, P<0.001), and histological grade ( P<0.001). High expression of Ki67 was associated with male ( χ2=17.33, P<0.01), smoking ( χ2=21.67, P<0.01), and histological grade ( P<0.001). Male ( HR=2.612, 95% CI: 1.173-5.815, P=0.019), smoking ( HR=2.651, 95% CI: 1.246-5.642, P=0.011), high pTNM stage ( HR=3.815, 95% CI: 1.792-8.122, P<0.001), high histological grade ( HR=5.277, 95% CI: 2.400-11.606, P<0.001), high P53 expression ( HR=5.950, 95% CI: 2.792-12.680, P<0.001), and high Ki67 expression ( HR=3.349, 95% CI: 1.554-7.221, P=0.002) were associated with poorer disease-free survival (DFS). Male ( HR=9.050, 95% CI: 1.113-73.586, P=0.039), smoking ( HR=8.428, 95% CI: 1.701-41.765, P=0.009), high histological grade ( HR=6.865, 95% CI: 1.756-26.834, P=0.006), high P53 expression ( HR=16.699, 95% CI: 3.369-82.761, P<0.001), and high Ki67 expression ( HR=7.558, 95% CI: 1.806-31.632, P=0.006) were associated with poorer overall survival. P53 high-expression was identified as an independent risk factor for both DFS ( HR=2.843, 95% CI: 1.192-6.778, P=0.018) and OS( HR=6.909, 95% CI: 1.202-39.720, P=0.030) in stage I INMA patients. The area under the time-dependent ROC curves for predicting 5-year overall survival after surgery were 0.738 for p53, 0.674 for Ki67, 0.638 for pTNM staging, and 0.587 for histological grade. Among these, p53 demonstrated the highest predictive efficacy. Conclusions:AI-assisted interpretation of P53 and Ki67 indices improves test result repeatability. With critical values of 62% and 20%, high P53 and Ki67 expression indicates poor prognosis, while high P53 expression is an independent risk factor for lower OS and DFS, serving as a reference for postoperative adjuvant therapy screening.
