Multicentre evaluation of the INTERPRET decision support system 2.0 for brain tumour classification

Abstract

In a previous study, we have shown the added value of (1) H MRS for the neuroradiological characterisation of adult human brain tumours. In that study, several methods of MRS analysis were used, and a software program, the International Network for Pattern Recognition of Tumours Using Magnetic Resonance Decision Support System 1.0 (INTERPRET DSS 1.0), with a short-TE classifier, provided the best results. Since then, the DSS evolved into a version 2.0 that contains an additional long-TE classifier. This study has two objectives. First, to determine whether clinicians with no experience of spectroscopy are comparable with spectroscopists in the use of the system, when only minimum training in the use of the system was given. Second, to assess whether or not a version with another TE is better than the initial version. We undertook a second study with the same cases and nine evaluators to assess whether the diagnostic accuracy of DSS 2.0 was comparable with the values obtained with DSS 1.0. In the second study, the analysis protocol was flexible in comparison with the first one to mimic a clinical environment. In the present study, on average, each case required 5.4 min by neuroradiologists and 9 min by spectroscopists for evaluation. Most classes and superclasses of tumours gave the same results as with DSS 1.0, except for astrocytomas of World Health Organization (WHO) grade III, in which performance measured as the area under the curve (AUC) decreased: AUC = 0.87 (0.72-1.02) with DSS 1.0 and AUC = 0.62 (0.55-0.70) with DSS 2.0. When analysing the performance of radiologists and spectroscopists with respect to DSS 1.0, the results were the same for most classes. Having data with two TEs instead of one did not affect the results of the evaluation.

Classification of brain tumours from MR spectra: the INTERPRET collaboration and its outcomes

“…A lesson learned from Reference 75 was that radiologists in public health system centres do not have the time to embark on formal evaluations of a DSS. Therefore, when it was necessary to check whether a new version of the system, with an added long‐ T E classifier, could improve on the prospective study, six radiologists with no particular knowledge of spectroscopy and three expert MR spectroscopists were left free to use the system to categorize the same 40 cases . In most classes and superclasses of brain tumours the AUCs were not significantly different from those of the first study , except that for A3s radiologists had an AUC DSS = 0.59 ( n = 54 cases/238 readings) whereas the expert spectroscopists reached an AUC DSS = 0.71 ( n = 27 cases/116 readings).…”

“…In principle, this does not seem to agree with previous literature on classifiers ; however, the exact use of the system was not recorded in the latter studies. One important difference is that in the early studies classifiers were developed using only cases from a few common brain tumour types , whereas in the two more recent evaluations the DSS was used on cases from 15 different WHO classes, some of them as uncommon as melanocytoma or brain LY .…”

Deep learning magnetic resonance spectroscopy fingerprints of brain tumours using quantum mechanically synthesised data

“…However, with the exception of the study conducted by Velido et al 16 , these are single‐centre studies, commonly performed on a single scanner, and have not been tested on independent patient populations. The proposed methodology used data acquired across multiple institutions 20–24 . CNN models were trained on the augmented synthetic spectra and tested on two independent datasets acquired by different scanners.…”

Evaluation of the INTERPRET decision-support system: can it improve the diagnostic value of magnetic resonance spectroscopy of the brain?

“…In the second version, the INTERPRET DSS 2.0, an option to use both short and long TEs, was introduced. This, however, did not improve the diagnostic accuracy, but the study showed that neuroradiologists inexperienced in use of MR spectroscopy could use the DSS with outcomes similar to those of experienced spectroscopists [ 3 ]. Another study demonstrated that the use of a combination of short and long TEs with the INTERPRET DSS could help in the differential diagnosis between glioblastoma multiforme and metastasis [ 4 ].…”