Can the computed tomography texture analysis of colorectal liver metastases predict the response to first-line cytotoxic chemotherapy?

BACKGROUND Artificial intelligence in radiology has the potential to assist with the diagnosis, prognostication and therapeutic response prediction of various cancers. A fewstudies have reported that texture analysis can be helpful in predicting theresponse to chemotherapy for colorectal liver metastases, however, the resultshave varied. Necrotic metastases were not clearly excluded in these studies and inmost studies the full range of texture analysis features were not evaluated. Thisstudy was designed to determine if the computed tomography (CT) textureanalysis results of non-necrotic colorectal liver metastases differ from previousreports. A larger range of texture features were also evaluated to identifypotential new biomarkers.AIMTo identify potential new imaging biomarkers with CT texture analysis which canpredict the response to first-line cytotoxic chemotherapy in non-necrotic colorectalliver metastases (CRLMs).METHODSPatients who presented with CRLMs from 2012 to 2020 were retrospectivelyselected on the institutional radiology information system of our privateradiology practice. The inclusion criteria were non-necrotic CRLMs with aminimum size of 10 mm (diagnosed on archived 1.25 mm portal venous phase CT scans) which were treated with standard first-line cytotoxic chemotherapy(FOLFOX, FOLFIRI, FOLFOXIRI, CAPE-OX, CAPE-IRI or capecitabine). The finalstudy cohort consisted of 29 patients. The treatment response of the CRLMs wasclassified according to the RECIST 1.1 criteria. By means of CT texture analysis,various first and second order texture features were extracted from a single nonnecrotictarget CRLM in each responding and non-responding patient. Associationsbetween features and response to chemotherapy were assessed by logisticregression models. The prognostic accuracy of selected features was evaluated byusing the area under the curve.RESULTSThere were 15 responders (partial response) and 14 non-responders (7 stable and 7with progressive disease). The responders presented with a higher number ofCRLMs (P = 0.05). In univariable analysis, eight texture features of the respondingCRLMs were associated with treatment response, but due to strong correlationsamong some of the features, only two features, namely minimum histogramgradient intensity and long run low grey level emphasis, were included in themultiple analysis. The area under the receiver operating characteristic curve of themultiple model was 0.80 (95%CI: 0.64 to 0.96), with a sensitivity of 0.73 (95%CI:0.48 to 0.89) and a specificity of 0.79 (95%CI: 0.52 to 0.92).CONCLUSIONEight first and second order texture features, but particularly minimum histogramgradient intensity and long run low grey level emphasis are significantlycorrelated with treatment response in non-necrotic CRLMs