Fractional flow reserve decision support system
A computed tomography (CT)-based clinical decision support system provides fractional flow reserve (FFR) decision support. The available data, such as the coronary CT data, is used to determine whether to dedicate resources to CT-FFR for a specific patient. A machine-learnt predictor or other model, with access to determinative patient information, is used to assist in a clinical decision regarding CT-FFR. This determination may be made prior to review by a radiologist and/or treating physician to assist decision making.
1. A method for fractional flow reserve (FFR) decision support using a computed tomography (CT)-based clinical decision support system, the method comprising:
scanning a patient with a CT system, the scanning providing coronary CT data representing a heart of the patient;
acquiring non-invasive patient data and biochemical measurements;
extracting values from the non-invasive patient data and the biochemical measurements for features of an input vector of a machine-learnt predictor of the CT-based clinical decision support system from the coronary CT data;
generating, by the machine-learnt predictor of the CT-based clinical decision support system based on the values for the features of the input vector, a clinical decision of whether to perform CT-FFR for the patient; and
transmitting the clinical decision.
2. The method of claim 1 wherein extracting, generating, and transmitting are performed by the CT system upon scanning and before a review by a treating physician of the patient.
3. The method of claim 1 wherein generating comprises determining no CT-FFR for the patient.
4. The method of claim 3 wherein generating comprises generating the clinical decision as invasive treatment of the patient or medicinal treatment without the CT-FFR.
5. The method of claim 1 wherein generating comprises predicting values for the CT-FFR without receiving input of segmentation.
6. The method of claim 1 wherein the machine-learnt predictor for whether to perform CT-FFR is based on training data with a ground truth from invasive FFR.
7. The method of claim 1 further comprising performing the scanning, extracting, generating, and transmitting as part of a hierarchal workflow with each of at least two stages determining whether to perform the CT-FFR, the first stage based on patient data other than the coronary CT data and the second stage being the scanning, extracting, generating and transmitting when the first stage indicates a decision other than performing the CT-FFR.
8. The method of claim 1 wherein extracting and generating are performed during performance of CT-FFR, and wherein the clinical decision is to cease the performance of the CT-FFR prior to completion.
9. The method of claim 1 wherein generating comprises generating the clinical decision as performance of the CT-FFR.
10. The method of claim 1 wherein transmitting comprises transmitting the clinical decision as information for physician decision.
11. The method of claim 1 wherein transmitting comprises displaying a decision tree showing the clinical decision, other possible decisions, and further treatment options resulting from the clinical decision.
12. A system for fractional flow reserve (FFR) decision support, the system comprising:
a computed tomography (CT) scanner for scanning a patient, the CT scanner configured to output coronary CT data for the patient; and
a decision support processor configured to acquire non-invasive patient data and biochemical measurements and to extract from the non-invasive patient data and the biochemical measurements an input feature vector from the coronary CT data, to apply the input feature vector to a machine-learnt classifier, and to output a clinical decision about performance of CT-FFR for the patient from the machine-learnt classifier in response to the application of the input feature vector.
13. The system of claim 12 wherein the decision support processor is configured to output the clinical decision prior to entry of an instruction about CT-FFR into a computerized patient record for the patient by a physician.
14. The system of claim 12 wherein the decision support processor is configured to output the clinical decision prior to review of the coronary CT data by a radiologist.
15. The system of claim 12 wherein the clinical decision comprises a recommendation to not perform non-invasive CT-based fractional flow reserve.
16. The system of claim 12 wherein the machine-learnt classifier is one of a cascade of machine-learnt classifiers, another of the machine-learnt classifiers in the cascade configured to output indication other than performance of the CT-FFR based on patient data other than the coronary CT data, wherein the one of the cascade is after the indication.
17. The system of claim 12 wherein the decision support processor is configured to apply and output during performance of CT-FFR, the clinical decision being whether to cease the performance of the CT-FFR.