A real-time prostate cancer detection technique using needle insertion force and patient-specific criteria during percutaneous intervention.

In this article, the authors present a novel real-time cancer detection technique by using needle insertion forces in conjunction with patient-specific criteria during percutaneous interventions. Needle insertion experiments and pathological analysis were performed for developing a computer-aided detection (CAD) model. Backward stepwise regression method was performed to identify the statistically significant patient-specific factors. A baseline force model was then developed using these significant factors. The threshold force model that estimated the lower bound of the cancerous tissue forces was formulated by adding an adjustable classifier to the baseline force model. Trade-off between sensitivity and specificity was obtained by varying the threshold value of the classifier, from which the receiver-operating characteristic (ROC) curve was generated. Sequential quadratic programming was used to optimize the CAD model by maximizing the area under the ROC curve (AUC) using a set of model-training patient data. When the CAD model was evaluated using an independent set of model-validation patient data, an AUC of 0.90 was achieved. The feasibility of cancer detection in real time during percutaneous interventions was established.