Page 25 - Delaware Medical Journal - November/December 2019
P. 25
PUBLIC HEALTH
from the conference’s project manager asking me about my credentials. Since space was limited and since many of the attendees were well-known in the oncology community, she undoubtedly wanted to know how I got on the list of those invited. I told her that I had been invited by Dr. Frelick. She replied, “Oh, OK” — such was
The conference had attracted enough attention that the science and medicine reporter for the New York Times, Gina Kolata, was in the audience. She later reported, as I recall, that the lung-cancer screening program showed promise. For various reasons, primarily because of the lack of a controlled trial, the ELCAP team was denied NCI funding. While the ELCAP study was of observational value, investigators were searching
for a larger randomized trial that had statistical validity. At the conference, the groundwork was then set for the later NCI trial in conjunction with the American College of Radiology — the National Lung Screening Trial (NLST).3 Dr. Rattay, in her superbly researched and presented article in the September/ Delaware Medical Journal, “Early Lung Cancer Screening in Delaware,” delves deeply into the results of the NLST and its implications for public health policy particularly for the state of Delaware. Missing in her discussion, and despite more than two decades of experience in screening for early lung cancer, are two critical issues that have yet to be adequately addressed.
1. OVERDIAGNOSIS
as “the diagnosis of a ‘cancer’ that would otherwise not go on to cause symptoms or death.”5 It is strikingly common. Welch estimates that in randomized trials there is overdiagnosis in 25% of mammographically detected breast cancers and in an astonishing 50% of
lung cancers detected by chest X-ray.5 This is understandable, as any disease- screening protocol casts a wide net that who may therefore suffer unintentional that constantly bedevils disease-screening investigators.
While overdiagnosis is barely mentioned in the NLST report, it is certainly there and no doubt blemishes the statistical results, particularly the published 20% reduction in lung cancer mortality.6 To tease it out of the NLST data, one must look at the in situ and minimally invasive adenocarcinomas (older term alveolar
cell carcinoma). In the NLST, there were 110 tumors (10.5%) that had the latter pathology.3 These lesions often present
on CT scans as ground glass opacities (GGOs); hazy lesions that do not obscure the underlying structures. Since they are very indolent tumors, they are almost always seen on the initial screening
scans. The tumor doubling time of GGOs is often several years and some in situ tumors remain unchanged for years.7 What
confounds investigators: which of the in situ tumors will progress and which will be overdiagnosed and overtreated?
2. INCIDENTALOMAS
These are common in any imaging procedure; more so in thin section
CT scans. In a Canadian lung cancer breast tumors.8 A majority of incidental follow-up, and/or specialist referral;
all with both added cost and ethical concerns. The radiologist faces dilemmas with medico-legal implications. What
in a screening study: the trace coronary lymph node, or the tiny adrenal nodule? Incidentalomas have rightly been called imaging’s Pandora’s box.
Rattay raises an important and controversial point when she writes:
Del Med J | November/December 2019 | Vol. 91 | No. 6
265
