PLoS ONE (Jan 2018)

Proposed method of histological separation between connective tissue disease-associated interstitial pneumonia and idiopathic interstitial pneumonias.

  • Mutsumi Ozasa,
  • Hiromi Ichikawa,
  • Shuntaro Sato,
  • Tomonori Tanaka,
  • Takeshi Johkoh,
  • Kensuke Kataoka,
  • Yasuhiko Yamano,
  • Yasuhiro Kondoh,
  • Hideki Nakamura,
  • Atsushi Kawakami,
  • Andrey Bychkov,
  • Hiroyuki Taniguchi,
  • Junya Fukuoka

DOI
https://doi.org/10.1371/journal.pone.0206186
Journal volume & issue
Vol. 13, no. 11
p. e0206186

Abstract

Read online

OBJECTIVES:Idiopathic interstitial pneumonia (IIP) and connective tissue disease -associated interstitial pneumonia (CTD-IP) are the two most common types of interstitial pneumonia. IIP and CTD-IP share common histological features, yet their clinical management is different. Separation of the two conditions based solely on histology can be challenging, and there are no established criteria. MATERIALS AND METHODS:We selected 105 consecutive cases of IIP (79 usual interstitial pneumonia and 26 non-specific interstitial pneumonia) and 49 cases of CTD-IP for derivation and 32 cases of IIP and 10 cases of CTD-IP for validation. Fourteen histological parameters were evaluated independently by two pathologists for derivation group and graded into 0 to 3. The association between the score for each marker and a diagnosis of CTD was investigated using Fisher's exact test and stepwise logistic regression analysis. A formula for calculating the probability of IIP and CTD-IP was constructed by the markers identified in the regression test with coefficients for each finding. The formula was confirmed using validation case group. RESULTS:Stepwise logistic regression analysis showed that plasmacytosis, lymphoid follicle with germinal center, and airspace fibrin were suggestive of CTD-IP and that fibroblastic foci, smooth muscle hyperplasia, cellular IP, dense perivascular collagen, and fat metaplasia were suggestive of IIP. The formula used to calculate the probabilities based on estimated values for each finding was created, and user-friendly web based app was composed at www.ctdip.com. On the validation study, 30 out of 32 IIP and eight out of 10 CTD-IPs were distinguished correctly by the app (Specificity: 93%, Sensitivity: 80%). CONCLUSIONS:We identified histological markers and derived a practical formula and user-friendly app to distinguish CTD-IPs from IIP.