KIMURA YuichiDepartment of Informatics Professor |
Japan has introduced Clinical Trials Act in April 2018 with unique definition of "clinical trials". Once defined as a clinical trial, the research has to be conducted by ones with medical doctor license. As a result, the act discourages medical and biomedical engineering researchers, mostly engineering-origin researchers without medical doctor license, to engage in research activities. In order to promote medical and biomedical engineering researches, Japanese Society for Medical and Biomedical Engineering issued guideline for researchers to check whether a particular research is classified as clinical trial under the clinical trial act definition on October 2019. Although the guideline provides general classification criteria, contributions of members to develop detailed criteria should be given for each research approach, such as electric-stimuli, motor-assistance, etc. is expected. Keywords: The Clinical Trials Act of Japan, Guideline, Classification Criteria.
Japan has introduced Clinical Trials Act in April 2018 with unique definition of "clinical trials". Once defined as a clinical trial, the research has to be conducted by ones with medical doctor license. As a result, the act discourages medical and biomedical engineering researchers, mostly engineering-origin researchers without medical doctor license, to engage in research activities. In order to promote medical and biomedical engineering researches, Japanese Society for Medical and Biomedical Engineering issued guideline for researchers to check whether a particular research is classified as clinical trial under the clinical trial act definition on October 2019. Although the guideline provides general classification criteria, contributions of members to develop detailed criteria should be given for each research approach, such as electric-stimuli, motor-assistance, etc. is expected. Keywords: The Clinical Trials Act of Japan, Guideline, Classification Criteria.
【目的】臨床的価値の高いFDG-PET/CT画像診断における医師の負担軽減及び診断精度向上のため、FDG-PET/CTの画像パターンに対する異常検知処理に基づく頸胸部病変の自動検出法を提案する。【手法】FDG-PETは、糖代謝亢進を反映し腫瘍部位では異常集積を呈する。そこで提案法では、CTの閾値処理及びモルフォロジ処理にて左右肺野を自動抽出し、肺野を基準に縦隔領域を自動抽出する。次に領域ごとの健常組織画素値に関する統計モデルを用いた異常検知による病変候補の検出を行う。異常検知に用いる統計モデルは健常画像の画素(HU, SUV)の2次元正規分布モデルとし、マハラノビス距離を異常度として計測する。左肺、右肺、縦隔で個別に算出した異常度の閾値処理にて病変候補を得る。得られた候補領域から微小体積のものを削除し、腫瘍が疑われる部位を検出する。図に病変検出例を示す。【結果】臨床撮像された食道癌、肺癌の10症例の全身PET/CTを用いて提案法の評価実験を行ったところ、検出感度94.7%(18/19病変)、過検出数14.0個/例の性能が得られ、提案法の有用性を確認した。【結論】今後の課題は、病変検出感度の改善や誤検出病変候補の削減処理の追加が挙げられる。
This study aims to develop a noise reduction algorithm for amyloid PET imaging. A visualization of a small amyloid-β (Aβ) deposit is required for early diagnosis of Alzheimer's disease (AD) pathology, and a noise in amyloid image should be reduced without any loss of spatial resolution. Conventional noise reduction algorithms use spatial averaging, but our proposed algorithm, CAKS, averages pixels that have similar kinetics of administered PET radiopharmaceutical. As a result of CAKS applying to actual clinical images, the noise is reduced and the image quality is improved. The contrast between gray and white matters is significantly improved (p<.05). We concluded that CAKS is applied to clinical amyloid PET imaging is useful for diagnosis of AD pathology.
A deposit of amyloid β (Aβ) to the central nervous system causes Alzheimer's disease (AD). A periodical evaluation on Aβ, a year, can therefor detect the beginning of AD, and it realizes an effective medical treatment for AD. For quantitative measurement of Aβ deposit is available using PET, and it requires to delineate the cerebellum gray matter as a reference region, when no Aβ deposit occurs physiologically and it has complicated shape. This study aim at proposing automated delineation algorithm based on a kinetics of administered PET tracer for Aβ imaging. The validation using a clinical data presented no significant difference (p < 0.05) between the amount of Aβ deposit using an ordinary manual delineation and the automated method. We can conclude that the algorithm is clinically useful.