BACKGROUND
Intrapancreatic accessory spleen (IPAS) shares similar imaging findings with hypervascular pancreatic neuroendocrine tumors (PNETs), which may lead to unnecessary surgery.
AIM
To investigate and compare the diagnostic performance of absolute apparent diffusion coefficient (ADC) and normalized ADC (lesion-to-spleen ADC ratios) in the differential diagnosis of IPAS from PNETs.
METHODS
A retrospective study consisting of 29 patients (16 PNET patients
vs
13 IPAS patients) who underwent preoperative contrast-enhanced magnetic resonance imaging together with diffusion-weighted imaging/ADC maps between January 2017 and July 2020 was performed. Two independent reviewers measured ADC on all lesions and spleens, and normalized ADC was calculated for further analysis. The receiver operating characteristics analysis was carried out for evaluating the diagnostic performance of both absolute ADC and normalized ADC values in the differential diagnosis between IPAS and PNETs by clarifying sensitivity, specificity, and accuracy. Inter-reader reliability for the two methods was evaluated.
RESULTS
IPAS had a significantly lower absolute ADC (0.931 ± 0.773 × 10
-3
mm
2
/s
vs
1.254 ± 0.219 × 10
-3
mm
2
/s) and normalized ADC value (1.154 ± 0.167
vs
1.591 ± 0.364) compared to PNET. A cutoff value of 1.046 × 10
-3
mm
2
/s for absolute ADC was associated with 81.25% sensitivity, 100% specificity, and 89.66% accuracy with an area under the curve of 0.94 (95% confidence interval: 0.8536-1.000) for the differential diagnosis of IPAS from PNET. Similarly, a cutoff value of 1.342 for normalized ADC was associated with 81.25% sensitivity, 92.31% specificity, and 86.21% accuracy with an area under the curve of 0.91 (95% confidence interval: 0.8080-1.000) for the differential diagnosis of IPAS from PNET. Both methods showed excellent inter-reader reliability with intraclass correlation coefficients for absolute ADC and ADC ratio being 0.968 and 0.976, respectively.
CONCLUSION
Both absolute ADC and normalized ADC values can facilitate the differentiation between IPAS and PNET.