Application of Artificial Intelligence in Analysis of Laboratory Data

1. 第一年的结果完成了血液检查分析方案，并与血液学专家合作，输入155名血液病患者的临床数据，以确定诊断的准确性。因此，有一些未被准确诊断的病例被Wintrobe的分类证实是由于其不可分类。因此，很明显，一些日本患者不符合Wintrobe的分类。另一方面，一些病例被错误诊断，这表明有必要纠正程序。然而，他们的程序已被证实具有显著的高诊断准确性。2. 第二年（最后一年）的结果对程序进行了修正，获得了几乎令人满意的诊断精度后，将其改写为微机程序。3. 研究计划的反思和未来扩展最初预期的生化测试数据分析计划规模比我们想象的要大。它的完成需要几年的时间。由于上述原因，原规划过于铺张，反映出研究的不完善。我们计划在未来几年完成实际应用的程序。

1. Results of the first year The programs for analysis of the blood tests was completed and the clinical data in 155 patients with blood diseases were input in cooperation with hematological specialists to establish the diagnostic accuracy. As a result, there are some cases which had not been diagnosed accuracy were confirmed to be due to their unclassifiableness by Wintrobe's classification. Therefore, it became clear that some Japanese patients are not in conformity with Wintrobe's classification. On the other hand, some cases were diagnosed erroneously, suggesting it necessary to correct the program. However, their program has been confirmed to have markedly high diagnostic accuracy. 2. Results of the second year(final year) After almost satisfactory diagnostic accuracy was obtained by correction ot the program, it was rewritten into the program for a microcomputer. 3. Reflection and future expansion of the research program The program of biochemical test data analysis anticipated at first was larger-scale than we imagined. Several years will be needed for its completion.For the above-mentioned reasons, the original plan was extravagant and the imperfectness of the study is reflected. We are planning to complete the program for practical use for several years ahead.