Source of abdominal pain identification in medical imaging
To assist a physician in diagnosis of trauma involving abdominal pain, scan data representing the patient is partitioned by organ and/or region. Separate machine-learnt classifiers are provided for each organ and/or region. The classifiers are trained to indicate a likelihood of cause of the pain. By outputting results from the collection of organ and/or regions specific classifiers, the likeliest causes and associated organs and/or regions may be used by the physician to speed, confirm, or guide diagnosis of the source of abdominal pain.
1. A method for identifying a source of abdominal pain, the method comprising:
scanning a patient with a computed tomography scanner, the scanning providing data representing an abdomen of the patient;
parsing, by an image processor, the data into a first portion representing a first organ and a second portion representing a second organ, the first and second organs being abdominal organs;
applying one or more first deep-learnt machine-trained classifiers to the first portion of the data by input of the data of the first portion to the one or more first deep-learnt machine-trained classifiers, the application resulting in output by the one or more first deep-learnt machine-trained classifiers of first likelihoods of multiple causes of pain for the first organ, the first deep-learnt machine-trained classifiers comprising first neural networks;
applying one or more second deep-learnt machine-trained classifiers to the second portion of the data, by input of the data of the second portion to the one or more second deep-learnt machine-trained classifiers the application resulting in output by the one or more second deep-learnt machine-trained classifiers of second likelihoods of multiple causes of pain for the second organ, the second deep-learnt machine-trained classifiers comprising second neural networks; and
generating an image of the patient from the data, the image including a plurality of the first and second likelihoods and the respective causes.
2. The method of claim 1 wherein scanning, parsing, applying the one or more first deep-learnt machine-trained classifiers, applying the one or more second deep-learnt machine-trained classifiers, and generating are performed during an emergency room visit by the patient.
3. The method of claim 1 wherein parsing comprises parsing with a third machine-learnt classifier.
4. The method of claim 3 wherein parsing comprises parsing the first portion with the third machine-learnt classifier and parsing the second portion with a fourth machine-learnt classifier.
5. The method of claim 1 wherein applying the one or more first deep-learnt machine-trained classifiers comprises applying just one first deep-learnt machine-trained classifier with the resulting first likelihoods of the multiple causes output by the just one first deep-learnt machine-trained classifier.
6. The method of claim 1 wherein applying the one or more first deep-learnt machine-trained classifiers comprises applying separate ones of the first deep-learnt machine-trained classifiers for each of the multiple causes.
7. The method of claim 1 wherein applying the one or more first deep-learnt machine-trained classifiers comprises outputting by the one or more first deep-learnt machine-trained classifiers the first likelihoods where the multiple causes comprise tumor, inflammation, stone, and bleeding.
8. The method of claim 1 wherein generating comprises generating the image with the plurality of the first and second likelihoods comprising a threshold limited number of the first and second likelihoods.
9. The method of claim 1 wherein generating comprises generating with values of the first and second likelihoods assigned to incremental ranges of at least 10% for each increment.
10. The method of claim 1 further comprising:
outputting clinical decision support with the image.
11. The method of claim 1 further comprising:
identifying a previous case in a database based on the first and/or second likelihoods; and
retrieving the previous case from the database.