Reply to Letter to the Editor re: Quantitative and qualitative assessment of non-contrast-enhanced pulmonary MR imaging for management of pulmonary nodules in 161 subjects

Abstract

Dear Sir, I read with keen interest the letter from Dr. Harders, and the original authors thank him for carefully reading our manuscript and making such constructive suggestions. Concerning the point about the smoking histories in the “Subjects” section, I mistook the Brinkman index for pack years and no one, including the authors and reviewers, noticed this error. When we calculate the Brinkman index as the number of cigarettes consumed per day multiplied by years of smoking, the cigarette smoke exposures of patients who smoked ranged from 120 to 3,600 (mean ± standard deviation of 830±720). However, when we recalculate as pack years, the cigarette smoke exposures of patients who smoked ranged from 6 (120/20) to 180 (3,600/20) (mean ± standard deviation of 41.5±36.0) pack years. Thank you for pointing this out to us and to other readers. According to the statistical analysis of qualitative assessment for distinguishing malignant nodules from benign nodules, we assessed the signal intensity of each nodule by using the five-point visual scoring system detailed in the manuscript. The letter to editor raised the folliowing point: “In the image analysis section much detail is given to describing a clinically relevant scoring system of 1–5, but in order to be further analyzed the data have to be dichotomized with a predecided cutoff (i.e. scores 1/2 = benign and scores 3–5 = malignant)”. However, signal intensities from muscle, nodules, fat and pulmonary parenchymal diseases can all yield different signal intensities according to the selected MR sequences. Thus in the previous literature, lesions including malignant and benign nodules, metastatic nodules, fibrosis and inflammation can all show increased or decreased T1 and/or T2 values [1–3]. Therefore, we discussed with our statisticians, and decided to compare the signal intensity of each nodule on each sequence with that of muscle by using the five-point visual scoring system similar to routine clinical practice and to determine the feasible threshold value by using ROC analyses. According to the results of ROC analyses, the feasible threshold values of qualitatively assessed T1WI, T2WI and STIR were determined as 2, 3 and 4, respectively. In addition, when these cutoff values were adapted, diagnostic capability of each sequence was optimised. Therefore, we used these cutoff values on T1WI, T2WI and STIR for distinguishing malignant nodules from benign nodules. This reply refers to the Letter to the Editor available at doi:10.1007/s00330-009-1439-x