Differentiating between benign and malignant ampullary strictures: a prediction model using a nomogram based on CT imaging and clinical findings

Machine learning for differentiating between pancreatobiliary-type and intestinal-type periampullary carcinomas based on CT imaging and clinical findings

PURPOSE

To develop a diagnostic model for distinguishing pancreatobiliary-type and intestinal-type periampullary adenocarcinomas using preoperative contrast-enhanced computed tomography (CT) findings combined with clinical characteristics.

METHODS

This retrospective study included 140 patients with periampullary adenocarcinoma who underwent preoperative enhanced CT, including pancreaticobiliary (N = 100) and intestinal (N = 40) types. They were randomly assigned to the training or internal validation set in an 8:2 ratio. Additionally, an independent external cohort of 28 patients was enrolled. Various CT features of the periampullary region were evaluated and data from clinical and laboratory tests were collected. Five machine learning classifiers were developed to identify the histologic type of periampullary adenocarcinoma, including logistic regression, random forest, multi-layer perceptron, light gradient boosting, and eXtreme gradient boosting (XGBoost).

RESULTS

All machine learning classifiers except multi-layer perceptron used achieved good performance in distinguishing pancreatobiliary-type and intestinal-type adenocarcinomas, with the area under the curve (AUC) ranging from 0.75 to 0.98. The AUC values of the XGBoost classifier in the training set, internal validation set and external validation set are 0.98, 0.89 and 0.84 respectively. The enhancement degree of tumor, the growth pattern of tumor, and carbohydrate antigen 19-9 were the most important factors in the model.

CONCLUSION

Machine learning models combining CT with clinical features can serve as a noninvasive tool to differentiate the histological subtypes of periampullary adenocarcinoma, in particular using the XGBoost classifier.

Differentiation of malignant from benign ampullary strictures: A prediction nomogram based on MR imaging and clinical findings

PURPOSE

To construct a predictive nomogram based on contrast-enhanced magnetic resonance imaging (MRI) and clinical findings for differentiating malignant from benign ampullary strictures.

METHOD

In this retrospective study, 76 patients with ampullary strictures (51 benign and 25 malignant) who underwent contrast-enhanced MRI were enrolled. Imaging findings were evaluated independently by two abdominal radiologists who reached consensus. Clinical findings were also collected. Significant findings for malignant ampullary strictures were assessed by univariable and multivariable logistic regression analyses. Based on the results of multivariable analysis, a nomogram to differentiate malignant from benign ampullary strictures was developed and internally validated.

RESULTS

In multivariable analysis, presence of an ampullary mass (odds ratio [OR]: 8.42, p = 0.047), bulging ampulla (OR: 8.32, p = 0.033), diffusion restriction of the ampulla (OR: 42.76, p = 0.004) on MRI, and jaundice (OR: 12.41, p = 0.019) were significant predictors of malignant ampullary strictures. A predictive nomogram was constructed using these findings. Among them, diffusion restriction of the ampulla showed the highest OR and predictor score on the nomogram. The calibration plots for internal validation achieved strong agreement between the predicted probabilities and the actual rates of malignant ampullary strictures.

CONCLUSION

A combination of significant contrast-enhanced MRI and clinical findings of ampullary mass, bulging ampulla, diffusion restriction of the ampulla, and jaundice may be useful in the prediction of malignant ampullary stricture.

Novel prognostic score based on the preoperative total bilirubin-albumin ratio and fibrinogen-albumin ratio in ampullary adenocarcinoma

BACKGROUND

The preoperative total bilirubin-albumin ratio (TBAR) and fibrinogen-albumin ratio (FAR) have been proven to be valuable prognostic factors in various cancers.

AIM

To detect the prognostic value of TBAR and FAR in ampullary adenocarcinoma (AC) patients who underwent curative pancreaticoduodenectomy.

METHODS

AC patients who underwent curative pancreaticoduodenectomy in the National Cancer Center of China between 1998 and 2020 were retrospectively reviewed. The prognostic cutoff values of TBAR and FAR were determined through the best survival separation model. Then, a novel prognostic score combining TBAR and FAR was calculated and validated through the logistic regression analysis and Cox regression analysis.

RESULTS

A total of 188 AC patients were enrolled in the current study. The best cutoff values of TBAR and FAR for predicting overall survival were 1.7943 and 0.1329, respectively. AC patients were divided into a TBAR-low group (score = 0) vs a TBAR-high group (score = 1) and a FAR-low group (score = 0) vs a FAR-high group (score = 1). The total score was calculated as a novel prognostic factor. Multivariable logistic regression analysis revealed that a high score was an independent protective factor for recurrence [score = 1 vs score = 0: Odds ratio (OR) = 0.517, P = 0.046; score = 2 vs score = 0 OR = 0.236, P = 0.038]. In addition, multivariable survival analysis also demonstrated that a high score was an independent protective factor in AC patients (score = 2 vs score = 0: Hazard ratio = 0.230, P = 0.046).

CONCLUSION

A novel prognostic score based on preoperative TBAR and FAR has been demonstrated to have good predictive power in AC patients who underwent curative pancreaticoduodenectomy. However, more studies with larger samples are needed to validate this conclusion.