Should CT play a greater role in preventing the resection of granulomas in the era of PET?

Benjamin J. May, Jeffrey M. Levsky, Alla Godelman, Vineet R. Jain, Brent P. Little, Panna S. Mahadevia, William B. Burton, Linda B. Haramati

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

OBJECTIVE. PET cannot distinguish between bronchogenic carcinoma and granuloma, but positive scans may prompt surgery. We systematically evaluated the CT appearance of resected carcinomas and granulomas to identify features that could be used to reduce granuloma resections. MATERIALS AND METHODS. We retrospectively identified 93 consecutive patients between January 2005 and November 2008 who had resection of a pulmonary nodule pathologically diagnosed as bronchogenic carcinoma or granuloma and preoperative imaging with CT and PET. Each nodule was evaluated on CT for size, doubling time, location, borders, shape, internal characteristics, calcification, clustering, air bronchograms, and cavitation. A diagnostic impression was rendered. Bivariate and logistic regression analyses were performed. Pre-PET data regarding the proportion of resected granulomas and carcinomas between January 1995 and December 1996 were reviewed. RESULTS. Sixty-eight percent (65/96) of nodules were carcinomas and 32% (31/96) were granulomas. The CT impression was benign in 65% (20/31) of granulomas and 5% (3/65) of carcinomas (p < 0.0001; negative predictive value [NPV], 87% [20/23]). Specific CT features significantly associated with granuloma were clustering, cavitation, irregular shape, lack of pleural tags, and solid attenuation. The combination of nonspiculated borders, irregular shape, and solid attenuation had an NPV of 86% (12/14). Granulomas represented 18% (9/50) of resected nodules in 1995 and 1996 (p = 0.066). CONCLUSION. CT findings reduce but cannot eliminate the possibility that a nodule is malignant. Outcomes-based clinical trials are needed to determine whether CT features of benignity can guide less-invasive initial management and reverse a concerning trend in granuloma resection.

Original languageEnglish (US)
Pages (from-to)795-800
Number of pages6
JournalAmerican Journal of Roentgenology
Volume196
Issue number4
DOIs
StatePublished - Apr 2011

Fingerprint

Granuloma
Carcinoma
Bronchogenic Carcinoma
Cluster Analysis
Logistic Models
Air
Regression Analysis
Clinical Trials
Lung

Keywords

  • Bronchogenic carcinoma
  • CT
  • Granuloma
  • PET
  • Pulmonary nodule

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Medicine(all)

Cite this

Should CT play a greater role in preventing the resection of granulomas in the era of PET? / May, Benjamin J.; Levsky, Jeffrey M.; Godelman, Alla; Jain, Vineet R.; Little, Brent P.; Mahadevia, Panna S.; Burton, William B.; Haramati, Linda B.

In: American Journal of Roentgenology, Vol. 196, No. 4, 04.2011, p. 795-800.

Research output: Contribution to journalArticle

@article{c2fc2ffe732b47668f1084d2a70eddec,
title = "Should CT play a greater role in preventing the resection of granulomas in the era of PET?",
abstract = "OBJECTIVE. PET cannot distinguish between bronchogenic carcinoma and granuloma, but positive scans may prompt surgery. We systematically evaluated the CT appearance of resected carcinomas and granulomas to identify features that could be used to reduce granuloma resections. MATERIALS AND METHODS. We retrospectively identified 93 consecutive patients between January 2005 and November 2008 who had resection of a pulmonary nodule pathologically diagnosed as bronchogenic carcinoma or granuloma and preoperative imaging with CT and PET. Each nodule was evaluated on CT for size, doubling time, location, borders, shape, internal characteristics, calcification, clustering, air bronchograms, and cavitation. A diagnostic impression was rendered. Bivariate and logistic regression analyses were performed. Pre-PET data regarding the proportion of resected granulomas and carcinomas between January 1995 and December 1996 were reviewed. RESULTS. Sixty-eight percent (65/96) of nodules were carcinomas and 32{\%} (31/96) were granulomas. The CT impression was benign in 65{\%} (20/31) of granulomas and 5{\%} (3/65) of carcinomas (p < 0.0001; negative predictive value [NPV], 87{\%} [20/23]). Specific CT features significantly associated with granuloma were clustering, cavitation, irregular shape, lack of pleural tags, and solid attenuation. The combination of nonspiculated borders, irregular shape, and solid attenuation had an NPV of 86{\%} (12/14). Granulomas represented 18{\%} (9/50) of resected nodules in 1995 and 1996 (p = 0.066). CONCLUSION. CT findings reduce but cannot eliminate the possibility that a nodule is malignant. Outcomes-based clinical trials are needed to determine whether CT features of benignity can guide less-invasive initial management and reverse a concerning trend in granuloma resection.",
keywords = "Bronchogenic carcinoma, CT, Granuloma, PET, Pulmonary nodule",
author = "May, {Benjamin J.} and Levsky, {Jeffrey M.} and Alla Godelman and Jain, {Vineet R.} and Little, {Brent P.} and Mahadevia, {Panna S.} and Burton, {William B.} and Haramati, {Linda B.}",
year = "2011",
month = "4",
doi = "10.2214/AJR.10.5190",
language = "English (US)",
volume = "196",
pages = "795--800",
journal = "American Journal of Roentgenology",
issn = "0361-803X",
publisher = "American Roentgen Ray Society",
number = "4",

}

TY - JOUR

T1 - Should CT play a greater role in preventing the resection of granulomas in the era of PET?

AU - May, Benjamin J.

AU - Levsky, Jeffrey M.

AU - Godelman, Alla

AU - Jain, Vineet R.

AU - Little, Brent P.

AU - Mahadevia, Panna S.

AU - Burton, William B.

AU - Haramati, Linda B.

PY - 2011/4

Y1 - 2011/4

N2 - OBJECTIVE. PET cannot distinguish between bronchogenic carcinoma and granuloma, but positive scans may prompt surgery. We systematically evaluated the CT appearance of resected carcinomas and granulomas to identify features that could be used to reduce granuloma resections. MATERIALS AND METHODS. We retrospectively identified 93 consecutive patients between January 2005 and November 2008 who had resection of a pulmonary nodule pathologically diagnosed as bronchogenic carcinoma or granuloma and preoperative imaging with CT and PET. Each nodule was evaluated on CT for size, doubling time, location, borders, shape, internal characteristics, calcification, clustering, air bronchograms, and cavitation. A diagnostic impression was rendered. Bivariate and logistic regression analyses were performed. Pre-PET data regarding the proportion of resected granulomas and carcinomas between January 1995 and December 1996 were reviewed. RESULTS. Sixty-eight percent (65/96) of nodules were carcinomas and 32% (31/96) were granulomas. The CT impression was benign in 65% (20/31) of granulomas and 5% (3/65) of carcinomas (p < 0.0001; negative predictive value [NPV], 87% [20/23]). Specific CT features significantly associated with granuloma were clustering, cavitation, irregular shape, lack of pleural tags, and solid attenuation. The combination of nonspiculated borders, irregular shape, and solid attenuation had an NPV of 86% (12/14). Granulomas represented 18% (9/50) of resected nodules in 1995 and 1996 (p = 0.066). CONCLUSION. CT findings reduce but cannot eliminate the possibility that a nodule is malignant. Outcomes-based clinical trials are needed to determine whether CT features of benignity can guide less-invasive initial management and reverse a concerning trend in granuloma resection.

AB - OBJECTIVE. PET cannot distinguish between bronchogenic carcinoma and granuloma, but positive scans may prompt surgery. We systematically evaluated the CT appearance of resected carcinomas and granulomas to identify features that could be used to reduce granuloma resections. MATERIALS AND METHODS. We retrospectively identified 93 consecutive patients between January 2005 and November 2008 who had resection of a pulmonary nodule pathologically diagnosed as bronchogenic carcinoma or granuloma and preoperative imaging with CT and PET. Each nodule was evaluated on CT for size, doubling time, location, borders, shape, internal characteristics, calcification, clustering, air bronchograms, and cavitation. A diagnostic impression was rendered. Bivariate and logistic regression analyses were performed. Pre-PET data regarding the proportion of resected granulomas and carcinomas between January 1995 and December 1996 were reviewed. RESULTS. Sixty-eight percent (65/96) of nodules were carcinomas and 32% (31/96) were granulomas. The CT impression was benign in 65% (20/31) of granulomas and 5% (3/65) of carcinomas (p < 0.0001; negative predictive value [NPV], 87% [20/23]). Specific CT features significantly associated with granuloma were clustering, cavitation, irregular shape, lack of pleural tags, and solid attenuation. The combination of nonspiculated borders, irregular shape, and solid attenuation had an NPV of 86% (12/14). Granulomas represented 18% (9/50) of resected nodules in 1995 and 1996 (p = 0.066). CONCLUSION. CT findings reduce but cannot eliminate the possibility that a nodule is malignant. Outcomes-based clinical trials are needed to determine whether CT features of benignity can guide less-invasive initial management and reverse a concerning trend in granuloma resection.

KW - Bronchogenic carcinoma

KW - CT

KW - Granuloma

KW - PET

KW - Pulmonary nodule

UR - http://www.scopus.com/inward/record.url?scp=79957874367&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79957874367&partnerID=8YFLogxK

U2 - 10.2214/AJR.10.5190

DO - 10.2214/AJR.10.5190

M3 - Article

C2 - 21427327

AN - SCOPUS:79957874367

VL - 196

SP - 795

EP - 800

JO - American Journal of Roentgenology

JF - American Journal of Roentgenology

SN - 0361-803X

IS - 4

ER -