Image-based Automated Counting of Plasma Cells and Marrow Cellularity in Core Marrow Biopsies

浆细胞（PC）通常负责制造大量抗体。当PC癌变时，它们会积聚在骨髓中，在那里它们会干扰正常血细胞的产生并破坏骨骼。最具侵袭性的PC癌是多发性骨髓瘤，占所有癌症的1%。其中最有价值的程序用于诊断这些癌症和确定肿瘤负荷是骨髓检查。该程序包括获得抽吸器和骨活检，病理学家检查显微镜准备并估计存在的PC的数量。这个数字目前是诊断PC癌和评估治疗反应的必要标准。然而，目前的显微计数方法缺乏精度。吸入物的计数受到血液污染的影响，基于骨组织准备的估计多少有些主观，充其量是半定量的。此外，总骨髓细胞量，这是有价值的评估许多血液学条件，是半定量评估。为了方便这些计数和提高准确性和精密度，我们开发了基于数字图像的软件来执行骨髓活检中PC和骨髓细胞的快速定量。该软件的初步评估提供了极好的结果，我们正在进行额外的测试以进一步验证它。用于PC的类似方法可以扩展到其他细胞类型的分析，如白血病细胞、淋巴细胞或其他谱系的细胞。

Plasma cells (PC) are normally responsible for manufacturing large amounts of antibodies. When PC become cancerous they accumulate in the bone marrow, where they interfere with the production of normal blood cells and destroy bone. The most aggressive form of PC cancers is called Multiple Myeloma, which constitutes 1% of all cancers. One of the most valuable procedures used in the diagnosis of these cancers and determination of tumor load is a bone marrow examination. This procedure consists of obtaining aspirates and bone biopsies where pathologists examine microscopic preparations and estimate the number of PC present. This number is currently a required criterion for diagnosis of PC cancers and also evaluation of response to therapy. However, current microscopic methods for the enumeration of PC lack precision. The counts in the aspirates are affected by blood contamination and estimations based on bone tissue preparations are somewhat subjective and at best semi-quantitative. Also, the total marrow cellularity, which is valuable in assessing numerous hematologic conditions, is evaluated semi-quantitatively. To facilitate these counts and improve accuracy and precision, we developed digital image-based software to perform a rapid quantitation of PC and marrow cellularity in bone marrow biopsies. The initial evaluation of this software provided excellent results and we are performing additional testing to further validating it. A similar approach used for PC could be extended to the analysis of other cell types such as leukemic cells, lymphocytes, or cells of other lineages. Although the testing and validations of the current software version demonstrated very good results, there are limitations with the program that preclude its universal use. For example, the software produces acceptable results as long as it is used with images obtained with a single microscope and a single camera. If either of these is changed, the software coding needs to be changed accordingly, a process that is not practical. Also, the application process is rather slow. For these reasons, the software language and algorithm are currently being modified not only to make the program flexible enough to accommodate different resolutions and other imaging conditions (colors, intensities, etc.) but also to make the software application faster and more efficient in order to make it more practical than the existing version. An updated version of the software has been tested but the advantages of this new version over the original version is only partially incremental. Improving the existing version further will necessitate additional and will be costly. Therefore, a manuscript with the original data as a proof of principle is being finalized and will be submitted for publication shortly.