Doctor of Philosophy
Hybrid PET/CT imaging is a newer cancer imaging modality which allows contemporaneous PET metabolic/molecular and CT anatomical imaging during a single diagnostic session on a single device. However, current PET imaging has relied heavily on measuring the tracer accumulation in the tumour as a surrogate of the pathologic activity targeted by the tracer while CT is limited to measuring tumour size or visualizing tumour enhancement pattern, all of which lack sensitivity and specificity in detecting cancer and assessing treatment response. Therefore, it is critical to develop pharmacokinetic techniques for assessing the physiologic/molecular characteristics of tumours with clinical PET/CT scanners to improve clinical decision made based on such imaging results. We hypothesized that quantitative pharmacokinetic and functional imaging using hybrid PET/CT could sensitively diagnose cancer and monitor its treatment response. My Ph.D. research focused on prostate and lung cancer. Patients with histologically confirmed cancer were evaluated using dynamic PET imaging with targeting tracers and CT Perfusion (CTP) with iodinated contrast agent to evaluate the metabolic and pathologic molecular activity and perfusion in tumours.
In the lung cancer study, 26 patients with early-stage non-small cell lung cancer (NSCLC) were evaluated with dynamic [18F]FDG and CTP before and after neoadjuvant stereotactic ablative radiotherapy (SABR) to assess imaging response and the results correlated with pathological evaluation. The most sensitive model to predict pathological complete response combined BVpre-SABR (baseline blood volume) from CTP and relative change in SUVmax from PET to yield sensitivity, specificity, positive (PPV) and negative (NPV) predictive value and area under receiver operating characteristic curve (AUC) of 0.85, 0.92, 0.92, 0.86 and 0.92 respectively.
In the prostate cancer study, 23 and 19 patients were evaluated with dynamic [18F]FCH/CTP or [18F]DCFPyL/CTP, respectively. The most sensitive parameter set to localize and detect prostate cancer is Ki (plasma net uptake rate) and k4 (dissociation rate constant) of [18F]DCFPyL with sensitivity, specificity, PPV, NPV and AUC of 0.95, 0.92, 0.70, 0.98 and 0.96 respectively with reference to the digital histopathological images.
In conclusion, our studies show that quantitative pharmacokinetic and functional parameters from dynamic PET and CTP can detect cancer and predict treatment response with acceptable performance metrics.
Summary for Lay Audience
Hybrid positron emission tomography (PET)/computed tomography (CT) imaging is a newer cancer imaging modality which allows better diagnosis and monitoring of treatment response of cancer. PET uses small amounts of radioactive tracers to evaluate organ and tissue functions whereas CT create detail anatomical image of the body. Combining CT with PET as in a hybrid PET/CT scanner will lend form to function.
However, current PET imaging is limited by subjective reading of images by experienced physicians while CT is limited to measuring tumour size or visualizing tumour shape. These qualitative features lack sensitivity and accuracy in detecting cancer and assessing treatment response.
Therefore, it is critical to develop quantitative (objective vs. subjective) techniques to assess the physiologic and molecular characteristics of tumours from clinical PET/CT images to inform clinical decision.
My Ph.D research focused on developing quantitative techniques to analyze prostate and lung cancer PET/CT images. In early-stage non-small cell lung cancer (NSCLC), PET images of glucose metabolism and CT blood volume images were able to predict cure of tumour as early as 8-week after stereotactic ablative radiotherapy (SABR). In prostate cancer, PET images of a tumour cell surface protein was very sensitive in identifying the most malignant tumour nodule so radiation can be concentrated on it to achieve a cure.
In conclusion, my project shows that quantitative analysis of PET/CT images can detect cancer and predict treatment response better than current subjective (non-quantitative) reading of these images.
Yang, Dae-Myoung (Danny), "Quantitative Dynamic PET & CT Perfusion for Cancer Imaging: Diagnosis & Monitoring Treatment Response" (2020). Electronic Thesis and Dissertation Repository. 7037.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.