Application Of Multi-Mode Imaging Based On Artificial Intelligence In Accurate Diagnosis Of Pancreatic Adenosquamous Carcinoma

Part 1:Logistic regression model based on CT for the precise diagnosis of pancreatic adenosquamous carcinomaPurpose:Pancreatic adenosquamous carcinoma(PASC)is a rare variant of pancreatic ductal adenocarcinoma(PDAC)with different biological behavior and treatment from PD AC,but PASC often is misdiagnosed as PD AC.At present,there is a lack of effective non-invasive methods to different PASC from PDAC.In this study,we developed and validated a radiomics model based on fully automatic segmentation of pancreatic tumors from enhanced computed tomography(CT)to different PASC from PDAC.Materials and Methods:In this retrospective study,patients with surgically resected histopathologically confirmed PASC and PDAC who underwent MRI scans between January 2011 and December 2020 were included in the study.According to time of treatment,they were divided into training and validation sets.Automated deep-learning-based artificial intelligence was used for pancreatic tumor segmentation.Logistic analysis was performed with conventional CT features,radiomic features and deep learning features to develop clinical,radiomics,and deep learning models in the training set.The models’ performances were determined from their discrimination and clinical utility.Results:This study included 332 patients with PDAC and 201 patients with PASC.The area under the ROC curve(AUC),sensitivity,specificity,and accuracy of deep learning model were 0.86(95%CI 0.82-0.90),75.00%,84.23%and 80.60%and those of clinical and radiomics were 0.81(95%CI 0.76-0.85),62.18%,85.89%,76.57%and 0.84(95%CI 0.80-0.88),73.08%,82.16%,78.59%in the training set.In the validation set,the area AUC,sensitivity,specificity,and accuracy of deep learning model were 0.76(95%CI 0.67-0.84),68.89%、78.02%and 75.00%,those of clinical and radiomics were 0.72(95%CI 0.63-0.81),77.78%,59.34%,65.44%and0.75(95%CI 0.66-0.84),86.67%,56.04%,66.18%.The DeLong test revealed significant differences between the models in the training sets.The DC As in the training and validation sets showed that if the threshold probabilities were>0.05 and>0.1,respectively,using the deep learning model to distinguish PASC from PDAC was more beneficial than the treat-all-patients as having PASC scheme or the treat-all-patients as having PDAC scheme.Conclusions:The deep learning model based on CT outperformed the clinical model and radiomics model to differentiate PASC from PDAC;it can,thus,provide important information for decision-making towards precise management and treatment of PASC.Part 2:Linear discriminant analysis based on MRI for the precise diagnosis of pancreatic adenosquamous carcinomaPurpose:Pancreatic adenosquamous carcinoma(PASC)is a rare variant of pancreatic ductal adenocarcinoma(PDAC)with different biological behavior and treatment from PD AC,but PASC often is misdiagnosed as PD AC.At present,there is a lack of effective non-invasive methods to different PASC from PDAC.In this study,we developed and validated a radiomics model based on fully automatic segmentation of pancreatic tumors from enhanced magnetic resonance imaging(MRI)to different PASC from PDAC.Methods:In this retrospective study,patients with surgically resected histopathologically confirmed PASC and PDAC who underwent MR scans between January 2011 and December 2020 were included in the study.According to time of treatment,they were divided into training and validation sets.Automated deep-learning-based artificial intelligence was used for pancreatic tumor segmentation.Linear discriminant analysis was performed with conventional MR features and radiomic features a to develop clinical,radiomics,and mixed models in the training set.The models’performances were determined from their discrimination and clinical utility.Results:Overall,389 and 123 patients with PDAC and PASC were included,respectively;they were split into the training and validation sets.The mixed model showed good performance in the training and validation sets(AUC:0.94 and 0.96,respectively).The sensitivity,specificity,and accuracy were 76.74%,93.38%,and 89.39%for the training set,respectively,and 67.57%,97.44%,and 90.26%for the validation set,respectively.The mixed model outperformed the clinical(P=0.001)and radiomics(P=0.04)models in the validation set.Conclusion:Our mixed model,which combined MRI and radiomic features,can be used to differentiate PASC from PDAC.