ESMO: Blood Test May Screen for Lung Cancer

ISTANBUL, Turkey, Oct. 3 -- A blood test may be able to distinguish between a malignancy in the lung and a benign pulmonary pathology, researchers reported here.

Using proteomic analysis of serum samples from 147 patients with lung cancer and 23 with other chronic pulmonary diseases, French investigators said they made an accurate determination 97% of the time.

By separating the proteins using mass spectrometry, they found that 31 out of 228 protein peaks differed significantly between the serum samples from patients with malignant and benign tissue, said William Jacot, M.D., of Hopital Arnaud de Villeneuve in Montpellier, and colleagues, at the European Society for Medical Oncology meeting.

Nineteen proteins increased in lung cancer and 12 decreased, relative to benign diseases Dr. Jacot said.

No single peak was adequate to classify the samples, however. Two peaks accurately distinguished the diseases with 87% accuracy. "While not so bad - better than conventional tumor markers - it is not enough to use in a screening setting," said Dr. Jacot. When they extended the test to include six peaks, the accuracy improved to 97%.

"The results are a promising improvement for screening," said Dr. Jacot. He cautioned that the approach is still in the early phase of development. The test needs to be validated still on an independent and larger set of patient samples.

Additionally, the current cancer patient population in the sample was recruited from the hospital population, and therefore included many stage III and IV patients. Stage I and II patients, who will be the target of an early screening technique, accounted for just 25% of the samples.

The cancer patient samples included both adenocarcinoma and squamous cell histology. Each were detected equally well by the test.

"This approach is highly important and if these peaks are validated in an independent set, it will have a tremendous effect on clinical procedures," commented Hans-Joachim Schmoll, M.D., of Martin Luther University in Halle-Wittenberg, Germany.

All patients included in the trial had been exposed to tobacco smoke and would thus be considered at high risk for lung cancer, and the targeted screening population.

If validated, the final design of the blood test will depend on how many protein peaks are required for high sensitivity and specificity. If it is a finite number, then the researchers can identify the proteins themselves and develop an antibody-based test that would be valuable in the clinic, Dr. Jacot said. If numerous peaks are required for accuracy, the final design would be more cumbersome,