Background: Immune checkpoint inhibitor-mediated insulin dependent diabetes (ICI-DM) is a rare and irreversible adverse event often presenting with life-threatening diabetic ketoacidosis (DKA). Pancreatic volume changes have been studied in classic Type 1 and Type 2 diabetes as a surrogate for functional β-cell mass. In this study, we investigate longitudinal changes in pancreatic volumes in patients who develop ICI-DM. Methods: Among patients with ICI-DM seen at our institution between 2014 and 2020, 20 patients who had serial CT scans of the abdomen and pelvis before and after ICI DM diagnosis were identified for inclusion in this study. Demographic data and clinical variables were obtained from the electronic medical record. Weight-adjusted pancreatic volumes were calculated from CT scans at three time points. The most recent CT scan prior to ICI initiation was used as baseline, the CT scan immediately prior to ICI DM diagnosis was used as a midpoint, and the most recent CT scan prior to the end of the study period was used as the final time point for comparison. Results: Median age was 63 years old (range 35 to 83). Renal cell carcinoma and melanoma were the most common cancer types. Male gender predominated (80%). 18/20 patients were on a PD-1 inhibitor with the remaining two on a PD-L1 inhibitor. After initiation of ICI therapy there was a variable response in pancreatic volumes prior to the diagnosis of ICI-DM with 20% patients experiencing a volume loss of > 10% and 25% experiencing a volume gain of > 10%. Volume loss nor volume gain prior to diabetes diagnosis was associated with presentation with DKA. 9/13 (69%) of patients who had pancreatic enzymes checked at diagnosis of ICI-DM had elevated levels. Pancreatic atrophy with a median volume loss of 41% was seen in all patients at a median of 14.9 months (range 3-77 months) after ICI-DM diagnosis. Most had more than 20% volume loss from baseline to most recent scan with no correlation in degree of volume loss with the time interval. There was no evidence of pancreatic ductal dilation, increased pancreatic fat nor any changes consistent with chronic pancreatitis. Conclusions: This study shows a variable response in pancreatic volumes after initiation of ICI in patients who progress to developing ICI-DM, though most had a significant decline in volume after the diagnosis of ICI-DM with long-term pancreatic atrophy. As β-cell mass is thought to comprise 1-2% of the pancreas, these findings may suggest both endocrine and exocrine compartment changes because of ICI-DM, though exocrine dysfunction has not been clinically described in this patient population. As these patients receive frequent imaging during treatment, fluctuations in pancreatic volumes with new or worsening hyperglycemia may portend the onset of ICI-DM and clinicians should have a low threshold to screen for this diagnosis as many will present with life-threatening DKA.