RESUMEN
BACKGROUND: The Gleason grading system is an important clinical practice for diagnosing prostate cancer in pathology images. However, this analysis results in significant variability among pathologists, hence creating possible negative clinical impacts. Artificial intelligence methods can be an important support for the pathologist, improving Gleason grade classifications. Consequently, our purpose is to construct and evaluate the potential of a Convolutional Neural Network (CNN) to classify Gleason patterns. METHODS: The methodology included 6982 image patches with cancer, extracted from radical prostatectomy specimens previously analyzed by an expert uropathologist. A CNN was constructed to accurately classify the corresponding Gleason. The evaluation was carried out by computing the corresponding 3 classes confusion matrix; thus, calculating the percentage of precision, sensitivity, and specificity, as well as the overall accuracy. Additionally, k-fold three-way cross-validation was performed to enhance evaluation, allowing better interpretation and avoiding possible bias. RESULTS: The overall accuracy reached 98% for the training and validation stage, and 94% for the test phase. Considering the test samples, the true positive ratio between pathologist and computer method was 85%, 93%, and 96% for specific Gleason patterns. Finally, precision, sensitivity, and specificity reached values up to 97%. CONCLUSION: The CNN model presented and evaluated has shown high accuracy for specifically pattern neighbors and critical Gleason patterns. The outcomes are in line and complement others in the literature. The promising results surpassed current inter-pathologist congruence in classical reports, evidencing the potential of this novel technology in daily clinical aspects.
RESUMEN
BACKGROUND: Previous studies have shown a good predictive power of the risk scoring method, Estimation of Physiologic Ability and Surgical Stress, in predicting mortality after open elective aortic aneurysm repair. The aim of the present study was to evaluate the physiologic component of this method to assess mortality risk in a different geographic population from previously published reports. METHODS: Operative, morbidity and mortality data were collected retrospectively from charts of patients submitted to elective open repair of an abdominal aortic aneurysm over an 8-year period. There were 214 patients, the median age was 70 (range: 48-91) years; 179 (83.6%) patients were men. The Preoperative Physiologic Risk Score (PRS), Surgical Stress Score, and Comprehensive Risk Score (CRS) values were categorized and compared with morbidity and mortality rates. RESULTS: There were 27 deaths (12.6%), and 81 (37.9%) patients experienced a postoperative complication that required medical intervention. There was a significant statistical difference for the values of PRS and CRS for patients who survived (0.53/0.63, respectively) and for those who died (0.88/1.02, respectively), p < 0.0001 for both values. There is a strong correlation between PRS and CRS values and development of complications (p < 0.0001). Surgical Stress Score did not correlate as strongly to development of complications (p = 0.0028). For PRS, the area under the receiver-operator characteristic curve was 0.844 (95% confidence interval: 0.747-0.941) for mortality and 0.725 (95% confidence interval: 0.650-0.799) for morbidity. For CRS, the area under the curve was 0.812 (95% confidence interval: 0.734-0.891) for mortality and 0.719 (95% confidence interval: 0.645-0.792) for morbidity. There was also a significant positive correlation between length of hospital stay and PRS and CRS scores (p < 0.0001). In this study, it was found that renal impairment has a significant positive correlation with mortality (p = 0.0008), with an odds ratio of 4.3. In a multivariate regression analysis, renal impairment failed to increase the accuracy of the model when associated with the other parameters considered in PRS. CONCLUSION: This study corroborates with the previous findings that the Estimation of Physiologic Ability and Surgical Stress model seems to be a promising method of predicting death and postoperative complications in patients undergoing open abdominal aortic aneurysm repair.
