Oops, you're using an old version of your browser so some of the features on this page may not be displaying properly.

MINIMAL Requirements: Google Chrome 24+Mozilla Firefox 20+Internet Explorer 11Opera 15–18Apple Safari 7SeaMonkey 2.15-2.23

Poster session 01

75P - Comprehensive assessment of 18F-FDG PET/CT images of cancer patients improves predictions of survival


10 Sep 2022


Poster session 01


Radiological Imaging;  Nuclear Medicine Imaging (Molecular Imaging);  Response Evaluation (RECIST Criteria)

Tumour Site

Lymphomas;  Head and Neck Cancers;  Thoracic Malignancies


Glenn Liu


Annals of Oncology (2022) 33 (suppl_7): S27-S54. 10.1016/annonc/annonc1037


G. Liu1, M. La Fontaine2, A. Weisman2, S.S. houshmandi3, O. Lokre2, R. Jeraj4, T. Perk2

Author affiliations

  • 1 Medicine And Medical Physics, University of Wisconsin Carbone Cancer Center, 53705 - Madison/US
  • 2 Research And Development, AIQ Solutions, 53717 - Madison/US
  • 3 Management, AIQ Solutions, 53717 - Madison/US
  • 4 Medical Physics, University of Wisconsin-Madison, 53705 - Madison/US

Abstract 75P


Standardized reporting of treatment response in oncology patients has traditionally relied on RECIST and PERCIST. While these endpoints are useful in prioritizing active drugs for further development, assessment of the limited number of lesions (up to 5) prevents a comprehensive evaluation of treatment response heterogeneity that most patients experience. The central hypothesis of our study was that a more complex evaluation of all lesions improves outcome prediction.


385 patients were investigated, consisting of 115 non-small cell lung cancer (NSCLC) patients, 142 head and neck cancer (HN) patients, and 128 diffuse large B-cell lymphoma (DLBCL) patients. These patients underwent chemoradiotherapy (NSCLC, HN) or chemotherapy (DLBCL), with 18F-FDG PET/CT scans at baseline and either post-treatment (NSCLC, HN) or during treatment (DLBCL). Comprehensive evaluation of all lesions between timepoints was performed using TRAQinform IQ technology (AIQ Solutions). 88 imaging features were extracted from each patient (including SUVmax, SUVpeak, total lesion glycolysis, and intra-patient heterogeneity features). The TRAQinform Profile was defined as the output of random survival forest models that predict overall survival in each cancer type. Random survival forest models were evaluated with 3-fold cross-validation. The predictive accuracy of the AIQ TRAQinform Profile was compared to fully automated RECIST and PERCIST values using Kaplan-Meier analysis.


TRAQinform Profile had a significant survival prediction in NSCLC (p<0.0001), HN (p<0.0001), and DLBCL (p=0.005). This contrasts with RECIST or PERCIST, which did not reach statistical significance (p>0.05) in NSCLC or DLBCL. Significant predictive power of RECIST and PERCIST was found in HN with p=0.0004 and p=0.007, respectively.


While RECIST and PERCIST have utility in drug development, they are insufficient at predicting survival in individual patients with metastatic cancer. Comprehensive assessment of all lesions, as opposed to the limited assessment recommended by RECIST and PERCIST, is necessary for accurate prediction of clinical outcomes.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

AIQ Global, Inc.


AIQ Global, Inc.


G. Liu: Financial Interests, Personal, Advisory Board: Janssen; Financial Interests, Personal, Ownership Interest, Co-founder AIQ Solutions: AIQ Solutions; Financial Interests, Institutional, Royalties: AIQ Solutions; Financial Interests, Institutional, Invited Speaker: Madison Vaccines. M. La Fontaine: Financial Interests, Full or part-time Employment: AIQ Global, Inc. A. Weisman: Financial Interests, Full or part-time Employment: AIQ Global, Inc. S.S. Houshmandi: Financial Interests, Full or part-time Employment: AIQ Global, Inc. O. Lokre: Financial Interests, Full or part-time Employment: AIQ Global, Inc. R. Jeraj: Financial Interests, Personal, Ownership Interest: AIQ Solutions. T. Perk: Financial Interests, Full or part-time Employment: AIQ Global, Inc.

This site uses cookies. Some of these cookies are essential, while others help us improve your experience by providing insights into how the site is being used.

For more detailed information on the cookies we use, please check our Privacy Policy.

Customise settings
  • Necessary cookies enable core functionality. The website cannot function properly without these cookies, and you can only disable them by changing your browser preferences.