CT Growth Assessment Reliably Diagnoses Solid Lung Nodules

NEW YORK—Computed tomography (CT) screening for lung cancer has revealed subtypes of nodules whose natural histories are being assessed in long-term follow-up, according to Claudia I. Henschke, PhD, MD, director, Division of Chest Imaging, and professor of radiology, Weill Medical College, Cornell University.

Classification is based on the extent to which the parenchyma within the nodule is obscured. "If it’s totally obscured, it’s solid. If it’s partially obscured, it’s part-solid, and otherwise, it’s nonsolid," she said at the 7th International Conference on Screening for Lung Cancer.

The first 1,000 baseline screens at her institution and New York University identified 233 nodules, she said. Of these 189 were solid, 16 part-solid, and 28 nonsolid. Among the 136 nodules that were 1 to 5 mm in size, 127 were solid, 2 part solid, and 7 nonsolid.

Cancer was detected in 27 cases, or 12% of the 233 nodules. The distribution included 14 solid nodules, 8 part-solid, and 5 nonsolid. No new cancers have been discovered in solid nodules in follow-up since the baseline screening in 1998, Dr. Henschke reported.

With solid nodules, she said, growth assessment within 3 months proved diagnostically reliable. "But that’s not quite the same for the part-solid and the non-solid lesions," she said. Five additional malignancies have been diagnosed in these subtypes, two in part-solid nodules and three in nonsolid.

In one patient, a high-resolution scan of a 15-mm part-solid nodule identified at baseline prompted recommendation of a biopsy of what was considered a stage I cancer. The patient declined and went to other institutions for other tests. "By the time there was clear evidence of growth, it was 4 years later," Dr. Hen-schke reported, "and at that point the tumor was stage IV." The cancer was identified as adenocarcinoma.

Another patient with a 10-mm part-solid nodule also initially refused to have a biopsy but had one 2 years later. The stage IA adenocarcinoma was resected.

The three nonsolid nodules that were eventually diagnosed as adenocarcinoma, mixed subtype, took 3 to 4 years to diagnose. Increases in nodule density and size were difficult to discern until the changes became marked, Dr. Henschke said. At baseline, these lesions were 23, 15, and 10 mm in size. Some other non-solid lesions continue to be followed, she said.

With the additional diagnoses, Dr. Henschke reported, crude malignancy rates for the nodule subtypes are now 7% for solid nodules, 63% for part-solid nodules, and 29% for nonsolid nodules.

The size of the malignant lesions differed by subtype, she noted, typically being about 3 mm for solid nodules, 10 mm for part-solid, and 7 mm for non-solid. Almost all of the part-solid nodules larger than 10 mm were malignant, she said.

In contrast to its value in solid nodules, growth assessment discriminates poorly between malignant and benign subsolid nodules, Dr. Henschke said. Perhaps, she added, current assessment techniques are inadequate and may lead to false-negative results. "I think a lot of work still needs to be done to determine which ones need to be biopsied and which do not," she concluded.

The subsequent series of 1,074 baseline CT screens found 242 nodules; 179 were solid, 10 part-solid, and 53 nonsolid. Since the same scanner with 10-mm slice thicknesses and 5-mm reconstruction increments was used as in the initial screening, the near doubling in nonsolid nodule detection (from 28 to 53), Dr. Henschke suggested, was due to increased awareness. "When we first started the screening," she said, "we weren’t as aware of these nonsolid nodules. We didn’t look at them so carefully."

For the third series of 936 patients, a scanner with 2.5-mm slice thickness was used, and 337 nodules were identified. The number of solid nodules 1 to 5 mm in size jumped to 280. All 9 part-solid nodules were 1 to 5 mm in size, as were 30 of the 48 nonsolid ones.

"Changing the scanner," Dr. Henschke commented, "is certainly going to increase the frequency of the smallest nodules that you find, whereas the rest of them stay fairly stable."

The increase in small nodules, she added, "need not lead to undue workups at baseline because we simply discount them and say ‘Come back in a year or 6 months, ideally in a year.’ But small nodules are very important. They can be used as relevant background information for interpreting the repeat scans."