2358. Can Detector-Based Dual-Energy CT Add Value In Characterizing Incidental Adrenal Lesions?
Authors * Denotes Presenting Author
  1. Jinjin Cao *; Massachusetts General Hospital
  2. Anushri Parakh; Massachusetts General Hospital
  3. Simon Lennartz; Massachusetts General Hospital
  4. Evita Joseph; Massachusetts General Hospital
  5. Avinash Kambadakone; Massachusetts General Hospital
To investigate the diagnostic performance of detector-based dual-energy computed tomography (dlDECT) in the evaluation of adrenal lesions.

Materials and Methods:
In this HIPAA-compliant, IRB-exempt retrospective analyses, 43 consecutive patients who underwent CT scans for suspected adrenal lesions on a dlDECT platform over a 16-month period (July 2018-November 2019) were identified. All scans were triphasic (TUE, PVP, 15-minute-delayed phases) examinations, where the PVP was acquired with DECT approach. VUE images were derived from the PVP acquisition. A radiologist, blinded to final diagnoses, quantified the attenuation for adrenal lesions and abdominal tissues on TUE, VUE, contrast and delayed phase images. Adrenal adenoma was diagnosed if CT density on both TUE and VUE was = 10 HU. The lesion absolute washout from TUE and VUE was calculated and compared. Lesion stability (over a 6-month period), biochemical analysis and in- and out-of-phase MR characteristics served as the reference standard for adrenal lesion characterization.

Forty-eight adrenal lesions were identified in 43 patients (27 female, 16 male; 61 ± 13 years; 81 ± 15kgs). The lesions measured 20.2 ± 10.2 mm (range: 8 - 50 mm). Based on the reference standard, 42 lesions were adrenal adenomas. The mean attenuation of all adrenal adenomas was higher on VUE compared to TUE (21.8 vs 5.5 HU; p<0.0001). The absolute difference in the attenuation of adrenal adenoma between the two images was 16.3 HU. The attenuation of liver parenchyma was also significantly higher on VUE (59.4 vs 53 HU; p<0.0001). The absolute differences in attenuation between VUE and TUE were larger for adrenal adenoma compared to liver parenchyma. Using a 10 HU cut-off value, 57.14% (24/42) lesions were characterized as an adenoma on TUE. This number reduced to 9.52 % (4/42) on VUE at the same attenuation cut-off value. Using an absolute washout (>60%) cut-off value, 76.19% (32/42) lesions based on TUE and 90.48% (38/42) lesions on VUE were characterized as an adenoma.

Quantitative attenuation measurements demonstrate variability in the between VUE and TUE images in the evaluation of incidental adrenal lesions.