IPMN high-risk stigmata over-call malignancy; higher cut-offs trade sensitivity for specificity

Design

Single-centre retrospective diagnostic accuracy study

Population

418 patients with surgically confirmed IPMN (173 malignant)

Primary outcome

Accuracy of high-risk stigmata on MRI for malignancy

Effect

Current criteria sensitivity 82.7%, specificity 66.9%; 6-point score AUC 0.771 to about 0.80 with adjusted cut-offs

In 418 patients who had surgery for pancreatic intraductal papillary mucinous neoplasm at one centre, of whom 173 had malignant lesions, researchers tested the international high-risk stigmata on preoperative MRI.

The current criteria had a sensitivity of 82.7% but a specificity of only 66.9%. Obstructive jaundice and suspicious cytology were almost always right (specificity 98.8-99.6%), but an enhancing mural nodule of 5 mm or more, or a main duct of 10 mm or more, had specificity of about 80%. Raising the cut-offs to about 12 mm for both pushed specificity to 82% but dropped sensitivity to 65%. A six-point score using adjusted thresholds performed slightly better overall (AUC 0.77 rising to about 0.80).

This does not change the guideline thresholds, and lower sensitivity means more missed cancers. What it supports is precise reporting: measure the nodule and the duct, state the measurements, and let the multidisciplinary team weigh borderline values against the patient's fitness for pancreatic surgery.

• Measure and report the enhancing mural nodule size and main duct diameter, not just whether a stigma is present.
• Flag jaundice-related or cytology-positive features, which were highly specific for malignancy.
• Expect a third of patients without malignancy to meet current high-risk criteria, based on this cohort.
• Discuss borderline nodules and ducts in a multidisciplinary meeting with the patient's surgical risk in mind.

Why it matters

Pancreatic resection carries real morbidity, and a third of benign lesions here met the criteria that point towards surgery.

Don't overread it

This is a single-centre surgical cohort; the adjusted cut-offs have not been validated elsewhere and current guideline thresholds are unchanged.

The statistics, in plain English

Specificity of 67% means a third of patients without malignancy still met high-risk criteria. Raising the cut-offs improves specificity but misses more cancers: there is no threshold that does both. The AUC improvement of 0.03-0.04 is modest.

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.