Abstract: A mechanism is provided to implement an artificial intelligence (AI) based alert mechanism system for alerting a medical professional of potential inaccuracies in medical image analysis. Responsive to receiving a medical image of a patient and a radiology report associated with the medical image, the AI based alert mechanism analyzes the radiology report to identify medical findings detected by the medical professional and analyzes the medical image to detect one or more medical findings associated with the medical image. Responsive to the AI based alert mechanism identifying one or more medical findings, the AI based alert mechanism compares the identified medical findings to those medical findings identified in the radiology report. Responsive to the AI based alert mechanism identifying a discrepancy between the identified medical findings to those in the radiology report, the AI based alert mechanism generates an alert to the medical professional who generated the radiology report.

Abstract: Systems and methods for determining an abnormality in an elongated structure in a three dimensional medical image. One system includes an electronic processor. The electronic processor is configured to determine a centerline of the elongated structure in the three dimensional medical image and determine a plurality of two dimensional cross sections of the three dimensional medical image based on the centerline. For each two dimensional cross section of the plurality of two dimensional cross sections, the electronic processor is configured to convert the two dimensional cross section to polar coordinates, fit a line to the elongated structure in the two dimensional cross section converted to polar coordinates, and reconvert the two dimensional cross section to Cartesian coordinates.

Abstract: Systems and methods for generating a training example to train artificial intelligence software to automatically determine a centerline of an elongated structure of three dimensional images. One system includes an electronic processor configured to receive a plurality of reference points for a subset of a plurality of slices of a first three dimensional image. Each of the plurality of reference points marks a centerline of the elongated structure within one of the subset of the plurality of slices. The electronic processor is configured to determine an order of the plurality of reference points and fit a spline curve to the plurality of reference points based on the order of the reference points to create the training example. The electronic processor is further configured use the training example to train the artificial intelligence software to automatically determine a centerline of an elongated structure in a second three dimensional medical image.

Abstract: Mechanisms are provided to implement a cognitive artificial intelligence training mechanism for simulating patients for developing artificial intelligence based medical solutions. The cognitive artificial intelligence training mechanism perturbs non-image based information of a real patient from a real patient data set forming perturbed non-image based information, The cognitive artificial intelligence training mechanism generates an artificial patient data in an artificial patient data set using the perturbed non-image based information and a non-perturbed medical image of the real patient. The cognitive artificial intelligence training mechanism then trains an operation of a learning algorithm utilized by the cognitive data processing system using real patient data in the real patient data set and the artificial patient data in the artificial patient data set.

Abstract: Mechanisms are provided to implement a cognitive artificial intelligence training mechanism for simulating patients for developing artificial intelligence based medical solutions. The cognitive artificial intelligence training mechanism perturbs non-image based information of a real patient from a real patient data set forming perturbed non-image based information. The cognitive artificial intelligence training mechanism generates an artificial patient data in an artificial patient data set using the perturbed non-image based information and a non-perturbed medical image of the real patient. The cognitive artificial intelligence training mechanism then trains an operation of a learning algorithm utilized by the cognitive data processing system using real patient data in the real patient data set and the artificial patient data in the artificial patient data set.

Abstract: An approach is provided to transform a first set of images retrieved from an annotated source image dataset. The transformation is based on image characteristics found in a model's domain, such as grayscale medical images. The first set of images can be common images unrelated to the model's domain. The approach pre-tunes the model by using the transformed images. The model is included in a question-answering (QA) system. The approach further trains the model using a second set of annotated images with the second set of images corresponding to the target domain, such as medical images. After training, a image, such as a medical image, is received at the QA system. The received image already has image characteristics of the target domain and no transformation is needed. The QA system responsively provides predictions pertaining to the received image.