Spectral CT separated intrapancreatic spleen from hypervascular tumours on one phase

Design

Retrospective diagnostic accuracy study, single-phase spectral CT

Population

85 patients with hypervascular pancreatic lesions (34 accessory spleens, 51 neoplasms)

Primary outcome

Discrimination of accessory spleen from neoplasm

Effect

Z_eff ratio AUC 0.969 vs 0.799 conventional; reader accuracy 88.1% vs 60.7%

A retrospective study included 85 patients with hypervascular pancreatic lesions on portal-venous phase spectral CT: 34 intrapancreatic accessory spleens and 51 neoplasms. Each lesion's density was compared with the spleen across conventional images, virtual monoenergetic images, iodine maps and effective atomic number maps.

Accessory spleens matched splenic values more closely on every series. Discrimination was highest on effective atomic number maps (AUC 0.969 overall, 0.992 in the pancreatic tail) and iodine density (AUC 0.960), against 0.799 for conventional images. Three radiologists grading tail lesions reached 88% accuracy with spectral images added, against 61% with conventional images alone, with better agreement between them.

An intrapancreatic spleen in the tail can look like a neuroendocrine tumour and lead to further imaging or even resection. Where spectral data are acquired, comparing the lesion with the spleen on iodine and atomic number maps is a quick check before recommending more work-up.

• For a hypervascular pancreatic tail lesion, compare it with the spleen on iodine and effective atomic number maps if spectral data exist.
• A close match to splenic values favours an intrapancreatic accessory spleen.
• Consider spectral reconstructions before recommending further imaging or biopsy for an indeterminate tail lesion.
• Confirm with established tests where doubt remains; this is a single retrospective study.

Why it matters

A confident accessory spleen call on routine CT can spare a patient further imaging, biopsy or surgery.

Don't overread it

Single-centre and retrospective — the thresholds need external validation.

The statistics, in plain English

An AUC of 0.97 means the measure ranked an accessory spleen correctly against a tumour about 97 times in 100. With only 85 patients from one centre, the true performance in other hands may be lower.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.