Estimation of prognosis of tongue cancer using tumor depth and margin shape obtained from ultrasonography

Sitti Fadhillah Oemar Mattalitti, Toshiyuki Kawazu, Shintaro Kawano, Tatsuya Ikari, Hiroko Wada, Kazunori Yoshiura

Research output: Contribution to journalArticlepeer-review

6 Citations (Scopus)

Abstract

Objectives: To evaluate the feasibility of using ultrasound-obtained tumor depth (TD) and margin shape (MS) to predict the prognosis of tongue cancer. Methods: Fifty-two Japanese patients with tongue cancer who underwent ultrasound examination between 2007 and 2012 were retrospectively evaluated. TD was measured at the deepest portion of the tumor. MS was classified as “pressure,” “wedge-shaped,” or “permeated.” Prognosis was assessed by local recurrence, lymph node metastasis, and overall survival rate. We classified the patients into a good prognosis group and poor prognosis group. Relationships among TD, MS, and patient prognosis were evaluated using Spearman rank correlation. A regression formula to predict prognosis using TD and MS was derived. Results: The correlation between TD and MS was significant (Spearman rank correlation ρ = 0.552, p < 0.01). Using a structure matrix, we identified the contribution of both predictors from an ultrasound image. The contribution of TD was 0.796, while that of MS was better at 0.906. The prognosis could be predicted using the following regression formula: D=-2.801+0.093x+0.872y, where D prognosis (risk probability), x = TD (mm), and y = MS (grade). The average value for good risk probability was −0.254 (standard deviation, 0.962), while that for poor risk probability was 0.566 (standard deviation, 1.08) (p < 0.05). The cut-off value for classifying cases was −0.00169. Using the formula above, 73.1 % of patients in the poor prognosis group were correctly classified. Conclusions: MS and TD are useful predictors of tongue cancer prognosis.

Original languageEnglish
Pages (from-to)101-107
Number of pages7
JournalOral Radiology
Volume33
Issue number2
DOIs
Publication statusPublished - May 1 2017

All Science Journal Classification (ASJC) codes

  • Dentistry (miscellaneous)
  • Radiology Nuclear Medicine and imaging

Fingerprint

Dive into the research topics of 'Estimation of prognosis of tongue cancer using tumor depth and margin shape obtained from ultrasonography'. Together they form a unique fingerprint.

Cite this