Virtual Library

Start Your Search

F. Cunha



Author of

  • +

    P1.18 - Poster Session 1 - Pathology (ID 175)

    • Event: WCLC 2013
    • Type: Poster Session
    • Track: Pathology
    • Presentations: 1
    • +

      P1.18-018 - Diagnosis of lung cancer: performance of pre-operative procedures in clinical practice correlated with characteristics of the tumor. (ID 3086)

      09:30 - 16:30  |  Author(s): F. Cunha

      • Abstract

      Background
      Recent treatments for advanced non-small cell carcinoma (NSCLC) have increased the demand for accurate diagnosis of NSCLC rendered by histology or cytology. However precise classification is not always possible

      Methods
      We investigated the performance characteristics of preresection procedures for diagnosis of NSCLC Database was searched for resected NSCLCs during 2008-2011 with corresponding preresection cytology and/or biopsy cases. The pre-resection diagnoses were correlated with resection diagnosis considering the type of bronchoscopic or transthoracic procedure, size, location of tumor in bronchoscopy and distance from thoracic wall in transthoracic cases. Pathologic data were reviewed by two pathologists and TAC by two pulmonologists. Pathology of resection was categorized as adenocarcinoma- AD, squamous cell- SQ, adenosquamous carcinoma (ADSQ), and large cell carcinoma-LCC. The bronchoscopy (BF) procedures- bronchial biopsy BB, distal biopsy DB, bronchial washing (BW), bronchial brushing (BBr), endobronchial/transbronchial biopsy (ETBX), and trans-thoracic procedures (TT)—transthoracic fineneedle aspiration(TT-FNA) and transthoracic needlecore biopsy(TT-NCB) were identified. Sensitivity of procedures was determined taking surgical pathology as reference; 95% confidence intervals were estimated by Wilson method. Agreement was evaluated using Cohen’sKappa. Univariate and multivariate logistic regression was used to evaluate factors possibly associated with absence of pre-surgery tumor diagnosis and lack of agreement in the subset with pre-surgical diagnosis.

      Results
      A total 189 patients were included, with 538 previous diagnostic procedures. The distribution for diagnosis was 105 AD, 48 SQ, 9 ADSQ, and 17 LC. Median (Max-Min) size of the tumor was 27mm (10-25). Median (Max-Min) distance from the thoracic wall was 7 mm (0 – 50). Procedures sensibility were: all 69% (63% - 76%), , TT 57% (46%-66,5%) , TT-NCB 71%, TT-FNA 33%,BF 54% (47%-62%)BB 63%, DB 54%, ETBX 32%, BBR 22%, BW 12%. The 131 patients in the subset with pre-surgery diagnosis agreement between pre and post-surgical diagnosis was: all procedures k=0,54, TT k=0,69, TT-NCB k=0,69, TT-FNA k=0,34, BF k=0,69, BB 0,59, DB k=0,73, ETBX k=0,80, BBR k=0,57, BW k=0,51. Concordance by histology was all procedures- AD 82%, SQ 78%; TT - AD85%, SQ 67%; NCB- AD 93%,SQ 75%; TT-FNA AD 78%, SQ 50%; BF AD 81%, SQ 87%; BB AD 82%,SQ 86; DB AD 100%, SQ 90%; ETBX 88% SQ 100%; BBR AD 79%, SQ 38%; BW AD 67%, SQ 60%. From factors possibly associated with diagnosis sensibility only the size of the tumor, in TT ODD 5% by additional mm (p=0,001), in BF ODD 3% (p=0, 0038) was significantly associated in univariated and multivariated analysis. From the factors possibly associated with agreement between pre-surgical and surgical diagnosis only the size of tumor in PATT was significantly associated in univariated and multivariated analysis.

      Conclusion
      All procedures had a substantial to almost perfect agreement in diagnosis of non-small-cell lung cancer subtype. Broncoscopy had 69% sensibility, lower than literature maybe because only operable tumors were considered, influenced by tumor size and an agreement in diagnosis independent of all factors considered. Transthoracic procedures had a sensibility lower than literature with substantial agreement influenced by tumor size. Broncoscopy was better in squamous-cell diagnosis and transthoracic procedures in adenocarcinoma maybe due to location in lung