The American College of Chest Physicians recommendations advocate that the first step in the evaluation of any nodule is to estimate the pretest probability of malignancy.2 In this setting we teach one of two approaches: clinical intuition or models using clinically available information to mathematically describe our intuition. The most frequently used model was published by Swensen et al3 in 1997, and though it was largely developed based upon nodules detected by chest radiography, it remains useful to this day, perhaps due to its simplicity.4 Despite the usefulness of these models, astute investigators continue to ask important questions such as whether we have mined the radiologic characteristics of nodules for all of their predictive ability. In this edition of CHEST (see page 464), Mehta et al5 publish a report that took the existing Swensen model and added volumetric nodule measurements to determine whether there is incremental predictive value in measuring lung nodule volume for detecting (or, conversely, excluding) cancer. The authors used the Swensen model to determine the probability of malignancy in 230 nodules (221 consecutive subjects). Clinical data and nodule characteristics were combined with volume measurements that were calculated using a commercially available product. Subjects were prospectively followed and a definitive diagnosis was established either by resection, biopsy specimen, or size stability for at least 2 years on serial CT scans. The investigators incorporated volume as a variable into the Swensen model in three separate ways: nodule volume (model 1), the ratio of volume to diameter (model 2), or sphericity index (model 3)—a metric defined as the ratio of measured nodule volume to the volume of a perfect sphere with a diameter equal to that of the nodule. Using a 0.5 probability of malignancy as a cutoff, they determined that models 1 (83%), 2 (88%), and 3 (88%) correctly classified more nodules than Swensen’s base model (67%). It should be emphasized that the main improvement in accuracy of the models incorporating volume measurements was the increased recognition of malignant nodules beyond what the Swensen model would predict. Volume measurements were, thus, better at pushing the pendulum toward cancer than toward benignity. Their inclusion of only subjects with nodules of < 15 mm diameter makes this particularly relevant to clinical practice, since larger nodules pose very little management dilemma.