我們回溯性分析自民國75年1月至79年12月間，新診斷的298例攝護腺癌患者；其中277位及206位者分別有手術前血清鹼性磷酸及攝護腺特定抗原值可供分析。我們將這兩個血清生化檢查與核醫骨骼掃瞄結果比較，發現血清鹼性磷酸對骨骼轉移病灶之陽性及陰性及陰性預值分別為80%及77%。當攝護腺特定抗原分異點設定在100ng/ml 時其陽性預測值為85%；然而在後續追蹤期間所有攝護腺特定抗原大於100ng/ml 的病患均出現轉移病灶。在攝護腺特定抗原小於4ng/ml 的病患中也有23%發現有轉移病灶。利用這兩個血清生化檢查來預測骨骼轉移與曾呸不夠準確；縱使兩者並用也不能完全取 骨骼核醫掃瞄。但是我們似乎可以建議只有對攝護腺特定抗原小於100ng/ml的病患才需要例行骨骼核醫掃瞄，如此一來可以省下44%的醫療成本而且不會遺漏任何一個骨骼轉移的病患。

From January 1986 to December 1990, a total of 298 patients with newly diagnosed adenocarcinoma of prostate were enrolled in this study. Serum alkaline phosphatase (Alk-P) and prostate specific antigen (PSA, Tandem-R: Hybritech) were available in 277 and 206 patients, respectively. The serum biochemical parameters were compared with the finding of Tc-99m MDP bone scan. Serm Alk-P alone had a positive predictive value of 80%, and a negative predictive value of 77%. When a cutoff value of 100 ng/ml was used, serum PSA had a positive predictive value of 85%; however, in the follow-up period, all the other scan-negative patients progressed to bone metastasis. Of the patients with normal serum PSA concentration (< 4 ng/ml), 23% of them had evidence of bone metastasis. Neither of these two serum markers was accurate enough to predict bone metastasis; even the combination of serum PSA and Alk-P did not have a sufficient accuaracy to replace routine bone scan. We may suggest that Tc-99m MDP bone scan is indicated for those who have a preoperative serum PSA concentration less than 100% ng/ml. On that basis, a patient with bone metastasis would be rarely missed, and about 44% of medical cost could be saved.