Hozumi Ikeda

Diagnosis of chronic liver disease from liver scintiscans by artificial neural networks

Artificial neural networks were used in the diagnosis of chronic liver disease based on liver scintiscanning. One hundred and thirty-seven patients with chronic liver disease (12 with chronic persistent hepatitis, 39 with chronic aggressive hepatitis, and 86 with cirrhosis) and 25 healthy controls were studied. Sixty-five subjects (10 healthy controls, 20 patients with chronic hepatitis, and 35 patients with cirrhosis of the liver) were used in the establishment of a neural network. Liver scintiscans were taken starting 20 min after the intravenous injection of 111 MBq of Tc-99m-phytate. The neural network was used to evaluate five items judged from information on liver scintiscans: the ratio of the sizes of the left and right lobes, splenomegaly, radioactivity in the bone marrow, deformity of the liver and distribution of radioactivity in the liver. The neural network was designed to distinguish between three liver conditions (healthy liver, chronic hepatitis and cirrhosis) on the basis of these five items. The diagnostic accuracy with the neural network was 86% for patients with chronic hepatitis and 93% for patients with cirrhosis. With conventional scoring, the accuracy was 77% for patients with chronic hepatitis and 87% for patients with cirrhosis. Our findings suggest that artificial neural networks may be useful for the diagnosis of chronic liver diseases from liver scintiscans.

ANALYSIS OF DIFFUSE PARENCHYMAL LIVER DISEASE BY LIVER SCINTIGRAMS: DIFFERENTIAL DIAGNOSIS USING FUZZY INFERENCE AND GENETIC ALGORITHM IN NUCLEAR MEDICINE

In nuclear medicine, diagnosis of diffuse parenchymal liver disease such as chronic hepatitis (CH) or liver cirrhosis (LC) is evaluated by size and distortion of the liver, distribution of radioactive tracer in the liver, size and activities of the tracer in the spleen, the degree of visualization of the bone marrow, etc. using colloid liver scintigraphy, It is not difficult to read a scintigram for a typical pattern; however in some cases it is difficult to distinguish between normal and CH or CH and LC visually. Therefore, we tried to use fuzzy inferences to perform differential diagnosis. Using fuzzy inference, differential diagnosis of LC could be performed up to 100%, but those of CH and Severe fibrosis (SF) could not be performed sufficiently. Therefore genetic algorithm was tried to determine the fuzzy rules. By combining genetic algorithm and fuzzy inference, CH, SF, and LC could be differentiated to the degree 70%, 60%, and 100%, respectively.

Relationship between Imaging Condition and Diagnostic Information in Bone Scintigrams.

Image quality of scintigrams is influenced by several imaging conditions. But, the relationship between image quality and diagnostic information has been not clearly evaluated. Therefore the diagnostic information was evaluated visually by radiologists using bone scintigrams changing the imaging conditions : total counts, image density, and distance from collimator to patient surface. The diagnostic information was considered the point of whether normal or abnormal scintigrams and in the abnormal scintigrams, the information of abnormal position, the abnormal area and the abnormal radioactive distribution on the bone scintigrams. Analysis of variance for the diagnostic information was performed. We concluded that using our methods the relationship between imaging condition and diagnostic informaton is analyzed quantitatively and the base of imaging technique is evaluated objectively.

Relationship between image quality and changes in spatial resolution for the gamma camera.

空間分解能の劣化と視覚的な画質との関係を定量的に検討した。コリメータと検体を密着した時のFWHMを基準にして, 離した時のFWHMとの差を△FWHMとすると, 画質上の変化を認識できる割合 (ρ) は, △FWHMが大きくなるに従いsigmoid curveを描いて大となった。Dendyらの方法およびファジィ理論を適用して, 画質の違いを充分認識できる最小△FWHMを検討すると, △FWHM=0.5mmとなった。この値はコリメータから約2cm離れた時の値に相当した。

(p, 5n) 反応によって製造した123Iを用いたSPECTに関する基礎的検討

The effects of higher-energy photon from 123I (p, 5n) on the SPECT image quality were evaluated. The quality was evaluated by image contrast and %rms. Image contrast had similar tendency to planar and SPECT FWHM value. %rms was affected by septal penetration. Using 140 keV high resolution collimator (140 keV HR), image contrast was superior to that for 300 keV medium energy collimator (300 keV ME), but septal penetration rate (SPR) was 18% and %rms was 10.5. When quantitation is required, the collimator with less SPR than 18% is recommended for SPECT imaging. Using 300 keV ME, SPR was 0.05%, but spatial resolution and image contrast were inferior to that for 140 keV HR.