Gerson S L, Talbot G H, Hurwitz S, Lusk E J, Strom B L, Cassileth P A
Am J Med. 1985 Jul;79(1):57-64. doi: 10.1016/0002-9343(85)90546-7.
Invasive pulmonary aspergillosis, a serious opportunistic infection in adult patients with acute leukemia, is difficult to diagnose antemortem. To identify patients with invasive pulmonary aspergillosis without reliance on invasive diagnostic procedures, a discriminant scorecard for invasive pulmonary aspergillosis based on clinical parameters was evaluated in a three-phase study. In phase I, the records of 62 patients, including 15 with invasive pulmonary aspergillosis, were reviewed. Eleven clinical parameters distinguished patients with invasive pulmonary aspergillosis from control subjects. These parameters were combined into a discriminant scorecard. In phase II, the discriminant scorecard was validated by a blinded, retrospective review of 94 consecutive admissions. The discriminant scorecard score was highly associated with the clinical outcome (p less than 0.0005). The sensitivity of the discriminant scorecard was calculated as a range from 62.9 to 92.8 percent and the specificity as a range from 87.5 to 98.3 percent. In phase III, the clinical utility of the discriminant scorecard was determined by its prospective application to 49 consecutive patient admissions. The discriminant scorecard identified patients with invasive pulmonary aspergillosis at an average of 4.1 days prior to clinical recognition of the disease and initiation of amphotericin B therapy. The discriminant scorecard outperformed a complex function based on multiple linear regressions, was easy to use, and did not require difficult calculations. Thus, for this patient population, the discriminant scorecard was an accurate, useful noninvasive screening test for invasive pulmonary aspergillosis. The scorecard allows more rapid clinical identification of patients with this infection and could lead to improved patient survival through earlier diagnostic and therapeutic intervention.
侵袭性肺曲霉病是成年急性白血病患者中一种严重的机会性感染,生前难以诊断。为了在不依赖侵入性诊断程序的情况下识别侵袭性肺曲霉病患者,在一项三阶段研究中评估了基于临床参数的侵袭性肺曲霉病判别计分卡。在第一阶段,回顾了62例患者的记录,其中包括15例侵袭性肺曲霉病患者。11项临床参数可将侵袭性肺曲霉病患者与对照受试者区分开来。这些参数被组合成一个判别计分卡。在第二阶段,通过对94例连续入院患者进行盲法回顾性研究,对判别计分卡进行了验证。判别计分卡得分与临床结局高度相关(p小于0.0005)。判别计分卡的敏感性计算范围为62.9%至92.8%,特异性计算范围为87.5%至98.3%。在第三阶段,通过将判别计分卡前瞻性应用于49例连续入院患者,确定了其临床实用性。判别计分卡平均在临床识别该疾病并开始使用两性霉素B治疗前4.1天识别出侵袭性肺曲霉病患者。判别计分卡优于基于多元线性回归的复杂函数,易于使用,且不需要复杂计算。因此,对于该患者群体,判别计分卡是一种准确、有用的侵袭性肺曲霉病非侵入性筛查试验。该计分卡能够更快速地临床识别这种感染的患者,并通过更早的诊断和治疗干预提高患者生存率。