Can the utility of V/Q scans for embolism be improved?

With suspected pulmonary embolism, the diagnosis is often critical: A missed clot can prove fatal, yet anticoagulation therapy is also risky. And diagnostic imaging is frequently unable to solve the dilemma, due to the low specificity of the dominant imaging exam for PE, ventilation-perfusion scintigraphy (V/Q).

A number of researchers are looking at alternatives to V/Q scans, particularly CT angiography, although a new meta-analysis in the Annals of Internal Medicine says the use of CT to diagnose PE hasn't been adequately evaluated to date (Ann Intern Med, February 1 2000, Vol. 132:3, pp. 240-242).

Meanwhile, a new position paper in RadioGraphics advocates recategorizing some of the indeterminate findings on ventilation-perfusion scans to improve their utility.

Radiologists have been interpreting V/Q scans as showing a high, moderate or low probability for years, generally using the criteria established for these categories by the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) study. But the low-probability category has been criticized for having an "unacceptably high prevalence of pulmonary embolism," note Dr. Paul Stein of Henry Ford Heart and Vascular Institute in Detroit and Dr. Alexander Gottschalk from Michigan State University in East Lansing.

In their paper, Stein and Gottschalk suggest that further stratification with a "very-low-probability" interpretation would help counter those concerns. The very-low-probability category would include abnormalities with a less than 10% likelihood of PE, including triple-matched defects in the upper or middle lung, nonsegmental perfusion abnormalities, the stripe sign, and three other findings (RadioGraphics, January 2000, Vol. 20:1, pp. 99-105,

Although the very-low-probability signs have been discussed individually in other publications, their compilation in RadioGraphics may help radiologists implement such interpretations, Gottschalk said in an interview. "I think they're not terribly difficult to use," noted Gottschalk, who has also been working to validate the criteria.

The benefit, Gottschalk believes, is that unlike low-probability readings, very-low-probability readings are no longer seen as nondiagnostic and can reassure clinicians about the truly low likelihood of PE.

"The hope is, given an appropriate clinical suspicion and a very-low reading, a patient will not be anticoagulated," Gottschalk said. Other nondiagnostic V/Q scans would appropriately be followed up with CT angiography, pulmonary angiography or serial ultrasound scans of the legs.

At Beth Israel Deaconess Medical Center in Boston, a majority of patients have non-diagnostic V/Q scans. Some 65% of patients require further testing to confirm or exclude PE, according to Beth Israel radiologist Dr. Max Rosen, who wrote about clinician attitudes toward misdiagnosis of PE in a recent issue of Academic Radiology (January 2000, Vol. 7:1, pp. 14-20).

If very-low-probability scans were considered to be "essentially normal" at his institution, then only 57% of V/Q scans would require follow-up testing, Rosen said.

But Beth Israel is also increasing its use of CT angiography for suspected pulmonary embolism. "I think the role of V/Q scanning will diminish as CT angiography becomes more widely accepted, specifically because a non-normal and non-high-probability result often provides little diagnostic information, and often necessitates additional testing," Rosen said.

Gottschalk believes V/Q scanning should eventually share the stage with CT angiography, in a stratification scheme based on the patient's initial chest x-ray. Patients with a clinical suspicion of PE but a normal or near-normal chest film would undergo V/Q, which is frequently definitive in such cases, Gottschalk said. Conversely, complicated chest films would be followed by CT angiography. The key to this approach, noted Gottschalk: "You have to be able to do both of these things (CTA and V/Q) well."

By Tracie L. Thompson staff writer
February 16, 2000

Copyright © 2000

Page 1 of 436
Next Page