Lung cancer is the most common malignant tumor and has become the main cause of cancer mortality. From a clinical point of view, it is important to predict tumor grading before treatment. Until now, several studies have explored the value of histogram analysis of apparent diffusion coefficient (ADC) for preoperative tumor grading in various malignant lesions including glioma, prostate cancer, and cervical cancer (1,2,3 ). The purpose of this study is to investigate the application of histogram analysis of ADC in determining the pathological grade of non-small cell lung cancer (NSCLC). Ethic committee of Kanazawa Medical University approved this retrospective study and waived to get informed consent. The study included fifty-four patients (32 men, 22 women; mean age 74.78 ± 7.07, range 56-88 years) with surgically diagnosed NSCLC. All MR images were obtained with a 1.5-T system (Magnetom Avanto, Siemens, Erlangen, Germany). Diffusion-weighted MR imaging (DWI) was performed in the axial plane using a spin-echo, echo-planar imaging sequence with respiratory triggering by navigator-echo method. The imaging parameters included: echo time=65 ms, b=0, 800 sec/mm2, field of view=350×230 mm, matrix=128×84, and section thickness=5 mm. Tumors with large cystic or hemorrhagic changes or with large consolidation component were excluded from the analysis. The ADC maps were generated and reviewed using Ziostation 2 (Ziosoft, Inc., Tokyo, Japan), and placed a 3-D volume-of-interest (VOI) on the tumor superimposing low b-value DWI. All ADC values within the VOI were used to capture the average ADC within the tumor. The ADC values were then binned to construct the histogram. The histogram parameters (mean, minimum, maximum, quartile, skewness, kurtosis, and percentiles of 5th-95th) were correlated with pathological grades (grade 1-3) using Jonkhere-Terpstra test. Comparison of histogram parameters between high grade (grade 3, n=15) and low grade (grades 1 and 2, n=39) was performed. For the parameters that were significantly different between high- and low-grade tumors, receiver operating characteristic (ROC) curve analysis was performed to calculate the sensitivity and specificity.Mean, maximum, quartile, 10th, 25th, 50th, 75th, 90th, and 95th percentiles of ADC showed significant differences among the pathological grades (p=0.000, 0.001, 0.001, 0.012, 0.002, 0.000, 0.000, 0.000, 0.000, respectively) (Table 1) . Mean, 50th, 75th, 90th, and 95th percentiles of ADC proved to be significant histogram parameters for differentiating high- from low-grade. ROC analyses of the histogram parameters between high- and low-grade showed that AUC (area under the curve) of the mean, 50th, 75th, 90th, and 95th percentiles of ADC were 0.706, 0.688, 0.713, 0.730, and 0.740, respectively. The 95th percentile of ADC achieved the highest AUC, with a cut-off value of 1634.1×10-6 mm2/sec and sensitivity and specificity of 0.846 and 0.667, respectively. Figure 1 shows the representative case.The 95th percentile of ADC is the most beneficial parameter for distinguishing high-from low-grade tumors. ADC histogram analysis on the basis of the entire tumor volume is useful for predicting the pathological grade of NSCLC.