doi:10.1016/j.meddos.2008.01.001 A BLAND-ALTMAN ANALYSIS OF THE BIAS BETWEEN COMPUTED TOMOGRAPHY AND ULTRASOUND PROSTATE VOLUME MEASUREMENTS AIME GLOI,PH.D., STEPHEN MCCOURT, M.S., CORRIE ZUGE, C.M.D., ANNDREA GOETTLER, C.M.D., SALLY SCHLISE, M.D., and GREG COOLEY, M.D. Department of Radiation Oncology, St. Vincent Hospital, Green Bay, WI (Received 29 January 2007; accepted 1 January 2008) Abstract—This study assesses the agreement between computed tomography (CT) measurements of prostate volume and those obtained by ultrasound (US), a well-established non-invasive technique. Twenty-six patients aged between 58 and 74 years were evaluated for prostate seed implant therapy using both CT and US measurements. The level of agreement between these 2 methods, which were strongly correlated (R 2  0.828; p < 0.0001), was determined through Bland-Altman analysis. The mean prostate volume ( one standard deviation) of the sample was 31.8  10.5 cc for the CT method and 27.0  8.2 cc for the US method. The prostate volumes obtained by CT were, on average, 17% larger than the corresponding volumes determined by US. The average bias between the 2 imaging methods is 4.80 cc or 15%, which is significantly larger than the clinically acceptable margin of 10%. © 2008 American Association of Medical Dosimetrists. Key Words: Bland-Altman, Computed tomography, Ultrasound, Seed implant, Volume measurement. INTRODUCTION According to the American Cancer Society, in the United States, prostate adenocarcinoma is now the most com- monly diagnosed cancer and the second leading cause of cancer death in men, surpassed only by lung cancer. The American Cancer Society estimates that 317,000 1 new cases will be diagnosed this year. Early diagnosis has become more common due to the widespread availability of prostate-specific antigen (PSA) screening, resulting in more patients being diagnosed with an early stage of this disease. 2,3 Various treatments are available, such as sur- gery, seed implantation, and external beam radiation therapy (EBRT). The choice of treatment depends on the patient, and is often based on several factors including age, general health, patient or physician preference, and the status of the cancer. One treatment that has seen recent widespread use due to improvements in imaging is brachytherapy, which delivers an appropriate dose of radiation directly to the tumor site via seed implantation. The dose to be deliv- ered by the seed is based on the estimated tumor volume. Although brachytherapy (usually with an Iodine-125 [ 125 I] seed implant) has frequently been used in conjunc- tion with EBRT; for certain kinds of tumors 4,5 some studies have reported that brachytherapy alone blocks prostate tumor growth in about 80% to 95% of patients. That achievement measures up favorably against the success rates of surgery and EBRT. In addition, brachy- therapy patients report fewer and milder side effects than patients receiving other treatments, who frequently suffer from incontinence, impotence, or rectal injury. 6 –12 In brachytherapy treatment, transrectal ultrasound (US) is typically used to measure the volume of the prostate before implantation of the seed. The benefits of US include real-time imaging during treatment of the prostate gland, and clear definition of its apex and cap- sule. On the other hand, ultrasound has difficulty pene- trating bone, and provides relatively poor spatial resolu- tion. It is also difficult to document US images in a standardized manner, and therefore to generate an ex- actly reproducible diagnosis. In short, the quality of an ultrasound volume determination is highly operator- dependent. Computed tomography (CT) is widely used in ra- diotherapy planning because of its low spatial distortion, and is routinely used in postoperative dosimetry for brachytherapy. Prostate volume measurements using CT are not as straightforward, however, as they are based on a contour drawn by the physician. A variation of this technique, contrast CT, is now customary in patients with prostate adenocarcinoma. The contrast provided by the intravesical medium turns out to be vital in distinguishing the base of the bladder from the prostate, and leads to better overall prostate visualiza- tion. As has been shown by several investigators, the main difference between US and contrast CT lies in their definition of the apex. 13,14 CT adds a margin of about 6 mm. As both methods are widely used as a basis for target volume delineation, the goal of this paper is to Reprint requests to: Aime Gloi, Ph.D., St. Vincent Hospital, Department of Radiation Oncology, 835 South Van Buren Street Green Bay, WI 54301. E-mail: agloi@stvgb.org Medical Dosimetry, Vol. 33, No. 3, pp. 234-238, 2008 Copyright © 2008 American Association of Medical Dosimetrists Printed in the USA. All rights reserved 0958-3947/08/$–see front matter 234