综合成像、生物标志物和临床检查结果有助评估不确定肺结节

圣迭戈——田纳西州纳什维尔范德比尔特英格拉姆癌症中心的Pierre P. Massion医生在美国癌症研究协会和国际肺癌研究协会主办的肺癌分子起源联合会议上指出，在通过CT筛查检出的结节中，约有30%符合不确定肺结节的标准。这些结节很少为癌症，因此如何对这类患者进行随访是个问题。Massion医生表示，临床上肺结节患者日益增加，随着筛查的进步，这个问题将愈发突出。通过非侵入方式早期准确区别良性病变和癌症具有重要意义。需加快实现早期诊断，因为手术可在最大程度上增加治愈的几率，但也需要减少因良性疾病进行开胸手术的例数。


Pierre P. Massion医生

8项大型肺癌筛查研究探讨了病变大小与肺癌概率之间的关系(Chest 2007;132[3 Suppl]:94S-107S)。直径＜5 mm、5~10 mm、11~20 mm和21~30 mm的病变是癌症的概率分别为0~1%、6%~28%、33%~60%和64%~82%。结节越大，其为癌症的概率越大。然而，事实上，大结节的数量非常少。不确定结节的直径为5~15 mm，这些结节是目前尚未找到最佳处置方案的结节。不确定肺结节为癌症的概率介于6%至60%，跨度较广。

结节形状可提供进一步信息。邻接裂隙和中心钙化的三角形通常提示良性病变，并且一般不需要随访。实性非钙化棘状结节为癌症的概率较高。部分实性结节非常令人担忧。这些结节最可能包含恶性肿瘤。非实性病变(也称为磨玻璃影)较为棘手且难以评估。它们代表20%的疾病概率。

小结节随时间推移的生长率很可能是最佳的成像标志物之一，但对于直径5 mm的小结节，体积分析的变异系数较大。

预测模型对于肺结节的评估至关重要，但现有预测模型得出的结果有30%是错误的。最好的3种预测模型来自梅奥医院研究(Arch. Intern. Med. 1997;157:849-55)、退伍军人事务部研究(Chest 2007;131:383-88)和PLCO(前列腺、肺、结直肠和卵巢)癌筛查试验(N. Engl. J. Med. 2013;368:728-36)。目前，ACCP 2013年肺结节评估指南推荐使用这些预测模型评估直径＞8 mm的结节(Chest 2013;143[5 Suppl]:e93S-120S)。目前尚无针对未曾吸烟者的模型，这是当前该领域的一大问题。Massion医生预测，血清生物标志物可能有助于决定哪些具有不确定肺结节的患者需要进行活检或切除，哪些患者只需随时间推移接受仔细观察即可。

一项整合了62个肺结节的临床、成像和蛋白生物标志物检查结果的研究显示，单纯依靠临床资料预测疾病的敏感性为50%(Cancer Epidemiol. Biomarkers Prev. 2012;21:786-92)。加用CT成像检查可使曲线下面积增至61%。增加检查血液中的生物标志物的话，可使曲线下面积增至69%。

Massion医生医生表示，整合门诊检查、成像检查和生物标志物检查结果的方法有助于改善这些结节的分类，是该领域发展的方向。Massion医生声明无相关经济利益冲突。

爱思唯尔版权所有  未经授权请勿转载

By: DOUG BRUNK, Internal Medicine News Digital Network

SAN DIEGO – About 30% of nodules detected by CT screening fit the criteria for an indeterminate pulmonary nodule. Very few of those nodules represent cancer, and the question is, what do you recommend for those patients in terms of follow-up?

"We’re encountering more and more patients with lung nodules in the clinic, and with the advance of screening, it will become even more of a problem. The numbers are tremendous," Dr. Pierre P. Massion stated at the Joint Conference on the Molecular Origins of Lung Cancer, sponsored by the American Association for Cancer Research and the International Association for the Study of Lung Cancer.
 
Dr. Massion, the Ingram Professor of Cancer Research at Vanderbilt-Ingram Cancer Center in Nashville, Tenn., said it’s important to differentiate – early, accurately, and noninvasively – benign lesions from cancer. "There is a race for early diagnosis, because surgery is the best chance for cure ... but we also need to decrease the number of thoracotomies performed for benign disease."

Data from eight large trials of lung cancer screening examined the relationship between lesion size and the probability of lung cancer (Chest 2007;132[3 Suppl]:94S-107S). The probability of cancer was 0-1% for lesions less than 5 mm in diameter; 6%-28% for those 5-10 mm, 33%-60% for those 11-20 mm, and 64%-82% for those 21-30 mm.

"The bigger the nodule, the greater the probability of cancer. In fact, however, the number of large nodules is very small," Dr. Massion said. "The indeterminate ones are between 5 and 15 mm in diameter, and these are the ones we struggle with how best to handle." The probability of cancer from indeterminate pulmonary nodules ranges from 6% to 60%, which is a large range.

The shape of the nodule provides additional information, Dr. Massion said. Triangular shape abutting a fissure and central calcification are generally indicators of benign disease and typically do not require follow-up. Alternatively, solid, noncalcified spiculated nodules have a high likelihood of being cancer. Part solid nodules are "very worrisome," he said. "These are most likely to contain malignancy. Nonsolid lesions, also called ground-glass opacities, are troublesome and difficult to assess. They represent about a 20% probability of disease."

The rate of growth of small nodules over time "is probably one of the best imaging markers, [but] for small nodules such as those 5 mm in diameter, the volumetric analysis has a large coefficient of variance," he said.

Prediction models are important to the evaluation of lung nodules, yet even with existing tools "we’re wrong about 30% of the time," he said. The best three prediction models come from studies of patients at the Mayo Clinic (Arch. Intern. Med. 1997;157:849-55) and the Veterans Affairs department (Chest 2007;131:383-88), and from patients enrolled in the PLCO (Prostate, Lung, Colorectal, and Ovarian) Cancer Screening Trial (N. Engl. J. Med. 2013;368:728-36). These prediction models are now recommended for use on nodules greater than 8 mm in diameter in the ACCP 2013 guidelines for evaluation of lung nodules (Chest 2013;143[5 Suppl]:e93S-120S).

"We have no models for never-smokers, which is a huge problem in the community at the moment."

Dr. Massion predicted that serum biomarkers might "come to the rescue" for deciding which patients with indeterminate pulmonary nodules might need to go for a biopsy or resection and which can be carefully watched over time.

In a separate study of 62 lung nodules that integrated clinical, imaging, and protein biomarker findings, clinical information alone resulted in about 50% sensitivity for predicting disease, "which is not great," said Dr. Massion, who was the principal investigator (Cancer Epidemiol. Biomarkers Prev. 2012;21:786-92). The addition of CT imaging increased the area under the curve to about 61%. Adding biomarkers in the blood raised the bar to about 69%.

"It’s not a panacea, but we show a trend toward improvement of classification of these nodules, which is where I think this field is going – integrating information from the clinic, imaging, and the discriminatory power of biomarkers."

Dr. Massion said that he had no relevant financial conflicts to disclose.