WASHINGTON—What is the role of imaging in diagnosing acute pancreatitis? A new study presented at DDW 2024 examined approaches to diagnosis without CT or MRI.
Related Content
In a prospective multicenter study from Boston, a novel diagnostic score using non-imaging parameters demonstrated excellent discriminatory accuracy for AP. “If operationalized, we estimate that 32% of early CT or MRI imaging could … be avoided,” said David X. Jin, MD, MPH, a gastroenterologist at the Center for Pancreatic Disease at Brigham and Women’s Hospital and an instructor at Harvard Medical School, in Boston.
Despite the majority of patients with AP fulfilling diagnostic criteria by having characteristic abdominal pain and serum lipase at least three times the upper limit of normal (ULN) at presentation, early imaging often is used for confirmation. “Some 90% of patients fulfill clinical criteria at presentation, and still 70% go on to have early diagnostic imaging. Why are we overutilizing early CT?” Dr. Jin noted.
Prospective Testing Of Diagnostic Score
Dr. Jin and his co-investigators sought to prospectively validate their AP diagnostic score at two academic centers based on 349 patients presenting to the emergency department with serum lipase at least three times the ULN (abstract 888). A research coordinator blinded to the final diagnosis assessed patients using the score’s eight non-imaging parameters related to patient history, symptoms and laboratory findings (Table). AP was established by expert review of full hospitalization records, including relevant imaging.
Of 349 patients, 157 (45%) ultimately were diagnosed with AP. Among the 192 patients without the disease (55%), the most common diagnoses were gastroenteritis, renal failure, alcohol intoxication, sequelae of recent surgery, acute hepatitis/decompensated cirrhosis and small bowel obstruction.
| Table. Diagnostic Scoring System For Acute Pancreatitis | |
| Clinical characteristic | Points |
|---|---|
| Previous AP episodes, n | 1a |
| History of cholelithiasis | 2 |
| Abdominal surgery in preceding 2 months | -2 |
| Epigastric pain | 2 |
| Worsening severity | 1 |
| Duration of pain =5 days | -1 |
| Pain severity, 10-point scale | |
| Mild (0-3) | 0 |
| Moderate (4-6) | 2 |
| Severe (7-10) | 3 |
| Serum lipase | |
| =3× to <10× ULN | 0 |
| =10× to =20× ULN | 1 |
| =20× ULN | 2 |
| a Points are given for each episode, to a maximum of 4 points. ULN, upper limit of normal. Based on DDW 2024 (abstract 888). | |
“Prospective scores demonstrated excellent prediction, with an area under the curve [AUC] of 0.91,” Dr. Jin reported. A score of at least 6 points achieved the highest diagnostic accuracy: an F-score of 82.0, which corresponded to a sensitivity of 81.5%, specificity of 85.9%, positive predictive value of 82.5% and negative predictive value of 85.1%. Only one serious alternative diagnosis (0.3% of the entire cohort) scored at least a 6.
Among the key differences between patients diagnosed with AP versus other disorders were:
- epigastric pain: 85% diagnosed with AP versus 19% with another disorder;
- severe pain: 71% diagnosed with AP versus 36% with another disorder; and
- mild serum lipase elevation (ULN of three times to less than 10 times): 27% diagnosed with AP versus 73% with another disorder.
Atypical Pain Patterns Require High Lipase Threshold
A lipase value higher than 500 U/L (more than eight times the ULN) can help diagnose AP in patients who have atypical abdominal pain, according to a study presented by investigators from Mayo Clinic Arizona at DDW 2024.
While a lipase level higher than three times the ULN with typical pain is sufficient to make a diagnosis in the absence of imaging, the cutoff for atypical abdominal pain is unclear.
“We believe the findings can help reduce the burden of imaging in the 8% to 15% of patients with atypical pain,” said Sarah Jahangir, MBBS, who was a research fellow at Mayo when the study was conducted but is now a resident in internal medicine at ECU Health, in North Carolina.
In their prospective study, Dr. Jahangir and her co-investigators evaluated 477 patients, analyzing the serum lipase cutoffs to determine the threshold above which AP could be reliably diagnosed without imaging (abstract 891). Patients were characterized as having typical pain (epigastric), atypical pain (diffuse/localized to one segment other than epigastric/nonacute/nonsevere), or no pain and then were given a diagnosis of AP or not.
A serum lipase level higher than 500 U/L accurately identified nine of 10 patients with AP on CT imaging, and accurately ruled out all 11 patients without AP on imaging. This result yielded a test sensitivity of 91% and specificity of 100%, Dr. Jahangir reported.
Among the 55 patients with atypical pain, the mean lipase level was 1,600 U/L in 19 patients with AP and 457 U/L in those with other diagnoses (P<0.0001). In 81 patients without pain and with serum lipase higher than three times the ULN, the mean lipase level was 751 U/L in those with AP and 601 U/L in those without it (P=0.34).
—C.H.
“Patients who present with acute pancreatitis were four times more likely to have pain localized to the epigastric region, two times more likely to describe the pain as severe, and three times less likely to have mild elevations in serum lipase,” he said.
In another study, another team of researchers examined atypical pain and lipase levels (see sidebar).
When to Use Imaging
“Of course, there are multiple reasons to get imaging other than for diagnostic confirmation, and these would still be completely valid,” Dr. Jin said. “But we focused on the very first CT or MRI. Studies have shown that out of all the imaging performed in AP, about half is done within the first 24 hours.”
In this study, imaging was performed within 24 hours of presentation in 227 patients (65%). Tests were ordered more often for patients predicted to have AP: 75% when scoring at least 6 points versus 57% when less than 6 points (P<0.001). When used early, imaging revealed an alternative diagnosis in only eight patients (7%) with scores of at least 6 and only one patient (1%) with a score of at least 8.
Dr. Jin acknowledged that since one-third of patients did not undergo early imaging and one-fourth had no imaging at all, the true yield of early imaging is unknown.
After the presentation, Phil Hart, MD, a pancreatic disease specialist at The Ohio State University Wexner Medical Center, in Columbus, had a question. “How confident were you with assigning the diagnosis of acute pancreatitis? For example, in a patient with a small bowel obstruction, without imaging can we definitively know there’s no acute pancreatitis on top of it? Maybe in a subanalysis, you could look at how the scoring system performed when you had complete information, which includes imaging.”
Although Dr. Hart agreed “with the direction here—we want to cut back on imaging,” he suggested to Dr. Jin that “as you’re developing this prediction model, I think it has to be held up to a gold standard.”
“That’s a valid question, but the problem is the lack of a gold standard,” Dr. Jin responded. “With a diagnosis that currently requires two out of three factors, by definition, there is no gold standard. Because of this, I think a lot of us still resort to imaging for confirmation.”
Drs. Hart and Jin reported no relevant financial disclosures.
This article is from the October 2024 print issue.