Abstracts

RETURN TO ABSTRACT LISTING


1462. Follow-Up Imaging Optimization of Patients on Clinical Trials for Stage 3 and 4 Lung Cancer
Authors * Denotes Presenting Author
  1. Erik Soloff *; University of Washington
  2. Michael Posch; University of Washington
  3. Priya Pathak; University of Washington
  4. Mladen Zecevic; University of Washington
  5. Anthony Sader; University of Washington
  6. Ryan O'Malley; University of Washington
  7. Carolyn Wang; University of Washington
Objective:
Clinical trial protocols for patients with lung cancer frequently require follow-up CT imaging after two cycles (six weeks) of conventional chemotherapy or two-three months of immunotherapy. Disease response per RECIST 1.1 guides treatment, with confirmation of progressive disease often resulting in treatment cessation. Routinely acquiring a follow-up CT chest, abdomen, and pelvis for all patients, regardless of baseline disease, may result in unnecessary radiation exposure and increased cost by imaging areas without metastasis. The purpose of our study was to determine whether locations of disease on baseline CT could predict sites of progression for patients with stage 3 or 4 lung cancer.

Materials and Methods:
Patients with stage 3 or 4 lung cancer that had disease progression while on a clinical trial were retrospectively identified through our tumor imaging response database. Locations of disease progression were recorded and categorized as occurring within the chest, outside the chest (abdomen, pelvis, and brain), or both inside and outside the chest. Disease burden at the time of baseline scan was also categorized in this manner. Predictive accuracy of baseline imaging in detecting future progression was summarized using sensitivity and specificity, while the feasibility of baseline was summarized using positive and negative predictive values.

Results:
A total of 180 patients with 205 imaging timepoints enrolled in 52 clinical trials were identified. Twenty-two patients developed progression of disease while on more than one clinical trial; 19 had progression in two different trials and 3 in three different trials. All cases (150 of 150) with disease progression in the chest had disease in the chest at baseline. 73 of 80 (91%) cases with disease progression in the abdomen had disease in the abdomen at baseline, while 13 of 21 (62%) cases with progression in the pelvis had disease in the pelvis at baseline. The positive predictive value of the CT pelvis at baseline was 27% (13/49) while the negative predictive value was 92% (92/100). Positive and negative predictive values of the abdominal CT at baseline were 70% (73/105) and 90% (63/70) respectively. Of 8 cases identified as disease-free in the pelvis on baseline CT, only 1 progressed exclusively in the pelvis. Similarly, of 7 cases that were disease-free on the baseline abdominal CT, only 2 progressed just in the abdomen. 252 of 269 (94%) scanned anatomic regions (chest, abdomen, pelvis, brain) with disease progression had disease in that region at baseline. If a body region was disease-free on the baseline scan, that region remained disease-free at the time of progression in 90% of cases.

Conclusion:
For patients with stage 3 and 4 lung cancer, anatomic location of disease on the baseline exam may be used to determine which body regions should be included on follow-up exams. Specifically, if no disease is detected in the pelvis at baseline, CT pelvis is unlikely to identify isolated progression on future follow-up imaging. Eliminating this region on follow-up scans will reduce patient radiation exposure and potentially reduce cost.