Search

A Correction to this paper has been published

Existing quantitative imaging biomarkers (QIBs) are associated with known biological tissue characteristics and follow a well-understood path of technical, biological and clinical validation before incorporation into clinical trials. In radiomics, novel data-driven processes extract numerous visually imperceptible statistical features from the imaging data with no a priori assumptions on their correlation with biological processes. The selection of relevant features (radiomic signature) and incorporation into clinical trials therefore requires additional considerations to ensure meaningful imaging endpoints. Also, the number of radiomic features tested means that power calculations would result in sample sizes impossible to achieve within clinical trials. This article examines how the process of standardising and validating data-driven imaging biomarkers differs from those based on biological associations. Radiomic signatures are best developed initially on datasets that represent diversity of acquisition protocols as well as diversity of disease and of normal findings, rather than within clinical trials with standardised and optimised protocols as this would risk the selection of radiomic features being linked to the imaging process rather than the pathology. Normalisation through discretisation and feature harmonisation are essential pre-processing steps. Biological correlation may be performed after the technical and clinical validity of a radiomic signature is established, but is not mandatory. Feature selection may be part of discovery within a radiomics-specific trial or represent exploratory endpoints within an established trial; a previously validated radiomic signature may even be used as a primary/secondary endpoint, particularly if associations are demonstrated with specific biological processes and pathways being targeted within clinical trials. KEY POINTS: • Data-driven processes like radiomics risk false discoveries due to high-dimensionality of the dataset compared to sample size, making adequate diversity of the data, cross-validation and external validation essential to mitigate the risks of spurious associations and overfitting. • Use of radiomic signatures within clinical trials requires multistep standardisation of image acquisition, image analysis and data mining processes. • Biological correlation may be established after clinical validation but is not mandatory. ispartof: EUROPEAN RADIOLOGY vol:31 issue:8 pages:6001-6012 ispartof: location:Germany status: published

BACKGROUND: Metastatic colorectal carcinoma (CRC) tumors with KRAS mutations in exon 2, codons 12 and 13, do not benefit from anti-epidermal growth factor receptor (EGFR) antibody therapy. Current best practice recommends that KRAS mutation analysis be included in patients with CRC as a prerequisite for treatment with cetuximab or panitumumab. We selected 50 CRC formalin fixed paraffin embedded (FFPE) samples to compare a qPCR assay (QIAGEN therascreen® KRAS RGQ PCR Kit) to droplet digital PCR (ddPCR). Wild type and six single nucleotide mutations in KRAS codons 12 and 13 were tested. Methods. DNA was extracted and used in assays employing primers paired with locked nucleic acid (LNA) or TaqMan probes. Droplets were generated and streamed through a droplet reader (BioRad Laboratories, CA). The QuantaSoft software (BioRad Laboratories) was used to analyze the ddPCR data. Results. In limit of detection studies, ddPCR detected 0.002 cpd of known mutant DNA in a background of 1 cpd of wt DNA. Specific KRAS mutations were detected in 16 (of 50) samples using LNA probes. Of the 16 samples tested by LNA probes, 13 were also tested using TaqMan probes, confirming results from the LNA assays. Results from ddPCR assays were compared to results from an outside commercial laboratory. The commercial lab detected and identified mutations in 10 of the 50 FFPE samples tested. ddPCR assays also detected mutations in 10 of these samples. An additional five samples were determined to be positive by commercial qPCR but no specific mutation was reported. ddPCR detected and identified specific mutations in three of these as well as in three samples not identified by commercial qPCR. Conclusions. ddPCR allele specific assays compared well to current commercially available KRAS mutation detection assays. Some cross reactivity was noted in ddPCR assays employing LNA probes but not TaqMan probes.

Medical imaging has seen substantial and rapid technical advances during the past decade, including advances in image acquisition devices, processing and analysis software, and agents to enhance specificity. Traditionally, medical imaging has defined anatomy, but increasingly newer, more advanced, imaging technologies provide biochemical and physiologic information based on both static and dynamic modalities. These advanced technologies are important not only for detecting disease but for characterizing and assessing change of disease with time or therapy. Because of the rapidity of these advances, research to determine the utility of quantitative imaging in either clinical research or clinical practice has not had time to mature. Methods to appropriately develop, assess, regulate, and reimburse must be established for these advanced technologies. Efficient and methodical processes that meet the needs of stakeholders in the biomedical research community, therapeutics developers, and health care delivery enterprises will ultimately benefit individual patients. To help address this, the authors formed a collaborative program-the Quantitative Imaging Biomarker Alliance. This program draws from the very successful precedent set by the Integrating the Healthcare Enterprise effort but is adapted to the needs of imaging science. Strategic guidance supporting the development, qualification, and deployment of quantitative imaging biomarkers will lead to improved standardization of imaging tests, proof of imaging test performance, and greater use of imaging to predict the biologic behavior of tissue and monitor therapy response. These, in turn, confer value to corporate stakeholders, providing incentives to bring new and innovative products to market.