Tumor assessment by Computed Tomography (CT) has become essential to oncologists in determining the therapeutic response of the metastatic tumor burden in cancer patients on therapeutic trials. This is done by measurement of a subset of "target" metastatic lesions on baseline and follow-up CTs. From these measurements a quantitative assessment of the change in tumor burden over time is calculated. Traditionally, radiologists' CT reports do not consistently include these measurements. One major reason is the time and effort needed to identify target tumors on follow-up scans and making measurements. Instead, the measurements are commonly obtained from images by the oncologists themselves or with the radiologist in a tumor measurement session/consultation.A recent survey we performed at our institution demonstrated that 1) oncologists spend an excessive amount of time making measurements or searching for measurements buried in our reports and matching them to the images in PACS, and 2) oncologists and radiologists prefer measurements in reports be hyperlinked to annotated images.In an effort to improve the content and utility of CT reports for oncologists, we embarked on a collaboration with our PACS vendor (Carestream Health, Rochester, NY) to explore the addition of capacities to PACS that would facilitate the generation of tumor measurement data by the radiologist and presentation in a Multi-media (MM) report with hyperlinks to images that would enable efficient use by oncologists. The recent PACS upgrade we helped develop (v 12.0 Vue PACS) includes the hyperlink capacity and tables and graphs as part of the report, and tools that facilitate lesion identification and quantification (temporal image registration, lesion segmentation and serial 3D localization of measurements; measurement on current exam is automatically related to measurements on prior exams).In this pilot study, we assessed the impact of the PACS upgrade on the timing of radiologists in generating and of oncologists in using MM reports on cancer patients.
RESULTSRadiologists' average dictation times were 11.9 (±5.6) and 12.6 (±4) minutes, before and after PACS upgrade, respectively. Although the reporting time has increased after PACS upgrade, the observed difference was not statistically significant in our study (P value = 0.53). This timing is on par with an average day on CT service in that it takes about 6 hours to dictate and measure 25 complex cancer follow up cases at our institution. Average time for an oncologist to assess tumor burden initially (text-only reports) was 15.4 (±5.9) minutes. Average time for oncologist to assess tumor burden using the multimedia reports was 6.2 (±2.9) minutes for a mean time savings of 8.9 minutes (Range 5 -14 minutes per study) when used the MM report to enter data into study forms (P<0.001).
CONCLUSIONOur pilot study results demonstrate that multimedia reports with data tables and hyperlinks to measurements on key images of target lesions facilitate analysis of tumor response by oncologists resulting in a significant time savings. These MM reports may be generated by radiologists without significant increase in reporting times. We anticipate in our facility alone up to 6-10 hours work by each team of oncologists and research staff could be saved using MM presentation. Although we measured time to complete the work by radiologists and oncologists, time is not the only domain that will be impacted by this innovation.With time saved and more straightforward presentation of data through quantitative MM reports, further studies in larger, more controlled settings can further test if MM reports are a more effective provision of care by improving oncologist's decisions and patient outcomes. Leveraging technology that provides professional-appearing content and media-rich reporting including links to images, reports, and the images may also become increasingly important to patients with the more widespread use of patient portals.
METHODSTwo radiologists recorded the times it took them to dictate CT exams (of the Chest, Abdomen and Pelvis) in which they measured target lesions on 20 consecutive cancer patients on therapeutic trials before and after the PACS upgrade. The difference between the average times was tested using permutation test.Additionally, an oncologist recorded the times to extract and tabulate target lesion measurements on 10 CT studies of five synovial sarcoma patients that had been reported prior to the PACS upgrade. The timing was also recorded for the same process on the MM reports of the same patients after these reports became available. The permutation test of paired samples was used to compare the mean time differences between extracting the needed data from traditional text only and MM reports by the oncologist.