Biomarker Science

MedQIA is a leader in bringing imaging biomarker science to clinical trials. Our biomarker R&D program is a multidisciplinary, collaborative effort encompassing imaging physics, computer vision, pattern classification, biostatistics, and radiological sciences.


In the setting of multicenter clinical trials, consistency and reproducibility of measurements are key elements. MedQIA achieves this through standardization of image acquisition parameters, image quality control metrics, and computer-aided image segmentation and analysis tools. MedQIA imaging biomarkers are developed using state of the art machine learning techniques applied to large research databases that correlate quantitative imaging features with biological changes and treatment outcomes.

Cutting-edge computer vision technology is used to extract quantitative biomarkers from images with greater reproducibility than subjective visual assessments. They provide early objective evidence of therapeutic efficacy for accelerated drug development. Additionally, predictive biomarkers can select patients who are most likely to benefit from targeted therapies, resulting in reduced clinical trial sizes and improved patient outcomes. The future of imaging biomarkers is in early, non-invasive decision making – “who should be treated, with what, and when.”

    • oncology
    • biomarkers

    MedQIA has developed a unique set of cutting-edge quantitative oncology biomarkers. These biomarkers are machine-learned from proprietary multi-modality databases that correlate early imaging features with treatment outcomes. They include prognostic, predictive and surrogate outcome biomarkers.

    Predictive imaging biomarkers non-invasively predict a patient’s response to a specific therapy; thus, they are useful for patient eligibility and treatment planning. Surrogate outcome biomarkers provide quantitative early detection of treatment effects or disease progression. These quantitative biomarkers are calculated by computer-aided analysis systems to maximize accuracy and reproducibility, and have tremendous potential to reduce the number of subjects and shorten clinical trials.

    The latest science is implemented in validated software according to MedQIA System Development Life Cycle (SDLC), and all computer systems are FDA 21 CFR Part 11 compliant, as applicable. MedQIA’s analyses span a full range of imaging modalities and organ systems. We provide Sub-Specialty Image Interpretation through a partnership with University faculty Radiologists and Oncologists.

    • lung
    • imaging
    • biomarkers

    MedQIA, one of the largest lung imaging core laboratories in the world, partners with university-based radiologists, oncologists and pulmonologists to conduct extensive research and development in Quantitative Image Analysis (QIA) of lung CT images. In particular the group has developed a FDA 510(k) cleared software toolkit of image analysis routines that forms the basis of MedQIA’s Imaging Biomarker Informatics System (IBIS). It employs a model-based engine to segment the lung, lobes, segments, and airways. IBIS provides computer-assisted diagnosis of the lung from CT for early diagnosis, treatment planning and outcome assessment.

    The QIA approach complements conventional pulmonary function tests (PFTs) because it is minimally-invasive and able to perform both global and regional assessment of the lung, thus providing greater sensitivity. IBIS allows a variety of validated quantitative measures to be derived from any segmented lung region or subregion including lung lobes and zones.


Oncology Biomarkers

Quantitative Tumor Assessment

MedQIA’s uses proprietary computer-aided tumor measurement technologies to maximize accuracy and reproducibility and reduce adjudication rates. MedQIA analysis tools are semi-automated, with computer generated 3D tumor boundaries that are then edited and approved by radiologists. After the tumor boundaries have been approved all measurements are calculated automatically. Tumor measurements support response assessment, characterization, and classification. The computed measurements include:

  • Size: diameters, volumetrics
  • Morphological: structural/density features, shape features, surface/margin features
  • Functional: radiotracer uptake (SUV), perfusion/permeability (DCE, DSC)

Lesion assessment tools are implemented within MedQIA’s FDA 510(k)-cleared IBIS information system and have been scientifically validated in peer-reviewed animal, phantom, and human studies. Response assessments are computed automatically and available criteria include: RECIST, WHO, MacDonald, RANO, PCWG2, PERCIST.

Bone Scan Lesion Assessment

MedQIA has developed a Bone Scan Lesion Area (BSLA) biomarker that can be computed semi-automatically from whole-body scintigraphic imaging as a measure of overall bone tumor burden. Development and validation, including correlation with outcomes, was performed in drug treatment trial cohorts with controls in subjects with metastatic castrate-resistant prostate cancer (CRPC).

MedQIA uses proprietary bone scan image analysis software to automatically detect and measure lesions. The system performs image normalization, anatomy-specific intensity thresholding, and classification to remove false positives. Results and presented to nuclear medicine physicians for review, editing, and approval. Then measures of tumor burden, including BSLA, are computed automatically.


Baseline BSLA and early changes post treatment have been found to be predictive of overall survival in patients with metastatic castration-resistant prostate cancer.

Lung Imaging Biomarkers

COPD Quantitative Assessment

MedQIA has an extensive array of imaging biomarkers for Chronic Obstructive Pulmonary Disease (COPD). These include novel biomarkers for treatment planning and outcome assessment of new minimally invasive lobar volume reduction techniques: lobar volume and density assessment, Fissure Integrity Score® (FIS), Target Lobe Atelectasis Score® (TLAS), and airway morphometry for bronchoscopic treatment planning.

Lung Fibrosis Scoring

MedQIA has developed proprietary software using machine-learning to classify lung texture patterns including fibrosis, honeycombing, and ground glass. This software was developed using texture feature extraction and machine learning. The Quantitative Lung Fibrosis (QLF) and Quantitative Interstitial Lung Disease (QILD) measures have been validated against outcomes and applied in assessment of interstitial and scleroderma lung disease.