RESUMO
PURPOSE: Artificial intelligence can reduce the time used by physicians on radiological assessments. For 18F-fluorodeoxyglucose-avid lymphomas, obtaining complete metabolic response (CMR) by end of treatment is prognostic. METHODS: Here, we present a deep learning-based algorithm for fully automated treatment response assessments according to the Lugano 2014 classification. The proposed four-stage method, trained on a multicountry clinical trial (ClinicalTrials.gov identifier: NCT01287741) and tested in three independent multicenter and multicountry test sets on different non-Hodgkin lymphoma subtypes and different lines of treatment (ClinicalTrials.gov identifiers: NCT02257567, NCT02500407, 20% holdout in ClinicalTrials.gov identifier: NCT01287741), outputs the detected lesions at baseline and follow-up to enable focused radiologist review. RESULTS: The method's response assessment achieved high agreement with the adjudicated radiologic responses (eg, agreement for overall response assessment of 93%, 87%, and 85% in ClinicalTrials.gov identifiers: NCT01287741, NCT02500407, and NCT02257567, respectively) similar to inter-radiologist agreement and was strongly prognostic of outcomes with a trend toward higher accuracy for death risk than adjudicated radiologic responses (hazard ratio for end of treatment by-model CMR of 0.123, 0.054, and 0.205 in ClinicalTrials.gov identifiers: NCT01287741, NCT02500407, and NCT02257567, compared with, respectively, 0.226, 0.292, and 0.272 for CMR by the adjudicated responses). Furthermore, a radiologist review of the algorithm's assessments was conducted. The radiologist median review time was 1.38 minutes/assessment, and no statistically significant differences were observed in the level of agreement of the radiologist with the model's response compared with the level of agreement of the radiologist with the adjudicated responses. CONCLUSION: These results suggest that the proposed method can be incorporated into radiologic response assessment workflows in cancer imaging for significant time savings and with performance similar to trained medical experts.
RESUMO
Glioblastoma multiforme (GBM) is the most common primary brain tumor in adults, and the limited available treatment options have not meaningfully impacted patient survival in the past decades. Such poor outcomes can be at least partly attributed to the inability of most drugs tested to cross the blood-brain barrier and reach all areas of the glioma. The objectives of these studies were to visualize and compare by matrix-assisted laser desorption ionization (MALDI) imaging mass spectrometry the brain and tumor distribution of the phosphatidylinositol 3-kinase (PI3K) inhibitors pictilisib (GDC-0941, 2-(1H-indazol-4-yl)-6-(4-methanesulfonyl-piperazin-1-ylmethyl)-4-morpholin-4-yl-thieno[3,2-d]pyrimidine) and GNE-317 [5-(6-(3-methoxyoxetan-3-yl)-7-methyl-4-morpholinothieno[3,2-d]pyrimidin-2-yl)pyrimidin-2-amine] in U87 and GS2 orthotopic models of GBM, models that exhibit differing blood-brain barrier characteristics. Following administration to tumor-bearing mice, pictilisib was readily detected within tumors of the contrast-enhancing U87 model whereas it was not located in tumors of the nonenhancing GS2 model. In both GBM models, pictilisib was not detected in the healthy brain. In contrast, GNE-317 was uniformly distributed throughout the brain in the U87 and GS2 models. MALDI imaging revealed also that the pictilisib signal varied regionally by up to 6-fold within the U87 tumors whereas GNE-317 intratumor levels were more homogeneous. Liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) analyses of the nontumored half of the brain showed pictilisib had brain-to-plasma ratios lower than 0.03 whereas they were greater than 1 for GNE-317, in agreement with their brain penetration properties. These results in orthotopic models representing either the contrast-enhancing or invasive areas of GBM clearly demonstrate the need for whole-brain distribution to potentially achieve long-term efficacy in GBM.