14 PERSPECTIVE 2019 Since a number of users of FIT in assessment of symptomatic patients see cancer detection as the most important role, there is considerable interest in using very low f-Hb as cut-offs. In addition, those wishing to use FIT as a rule-out test to reassure those patients who are unlikely to have significant bowel disease, are also interested in using very low f-Hb cut-offs, so as to minimise the chances of missing disease. So, the question for both approaches, is “how low can FIT go?” Detectability Characteristics – LoB, LoD and LoQ This question can only be objectively answered by consideration of what are correctly termed the “detectability characteristics”. These are particularly relevant for those examinations, like FIT, that focus on the use of low analyte concentrations. If samples of faeces which had no haemoglobin present were analysed, a spectrum of results (analytical “signals”) would be found. The maximum f-Hb found on analysis of haemoglobin free samples is called the “Limit of Blank”, often denoted as the LoB. The LoB is defined as the highest measured result likely to be observed (typically at 95% probability) for a sample containing no f-Hb (a blank sample). If samples with very low f-Hb, that is samples with f-Hb slightly higher than the LoB, were analysed, the results (analytical signals) would be generally higher than the LoB. Now, when these f-Hb results became clearly significantly different from the LoB, the “Limit of Detection”, the LoD, has then been exceeded. The LoD is defined as the lowest f-Hb likely to be reliably distinguishable from the intrinsic analytical “noise”, the “signal” produced in the absence of analyte (blank), and at which detection becomes feasible. LoD is calculated as LoB + 1.645 x SD of samples with low f-Hb. The factor 1.645 is the “one-sided” number of analytical standard deviations (SD) that gives 95% probability. The LoD will be higher than the LoB. If samples with f-Hb equal to or higher than the LoD were analysed, the “Limit of Quantitation”, the LoQ, might be exceeded. The LoQ is defined as the lowest f-Hb at which the analyte can not only be reliably detected, but at which some predefined goals for analytical performance are met. In consequence, this definition means that analytical performance specifications (APS) must be decided before the LoQ can be determined and applied. Analytical Performance Specifications The current international consensus is that setting APS can be performed using one of three models: Model 1: Based on the effect of examination performance on clinical outcomes: Model 2: Based on components of biological variation of the measurand: Model 3: Based on the state of the art. At this point in time, data are available only to facilitate a Model 3 approach. It has been proposed that the interim APS for analytical imprecision (and measurement uncertainty) is CV ≤ 10%. This should be lowered when and if FIT technology and methodology improve. It is vital to note that, while LoB and LoD are basically statistical concepts, LoQ depends on the defined APS. A typical “imprecision profile”, in which coefficient of variation (CV, %) is plotted against the f-Hb is shown in Figure 1. Faecal immunochemical tests for haemoglobin (FIT) are now widely used in asymptomatic screening for bowel cancer and in assessment of patients presenting with lower bowel symptoms. FIT might also be of value in other clinical settings such as surveillance of patients with previous bowel disease. Quantitative FIT provide estimates of the faecal haemoglobin concentration (f-Hb) for use in all of these. It is now well documented that the f-Hb is related to the severity of bowel disease. Thus, lowering the f-Hb cutoff used to trigger clinical action confers better detection of bowel cancer, adenoma and other serious bowel pathology. The clinical sensitivity increases, but the positivity rises, the demand for colonoscopy increases, the number of false positive results rises and the positive predictive value falls. Normal Low risk adenoma High risk adenoma Cancer Faecal Haemoglobin How Low Can FIT Go? Professor Callum G. Fraser Explores the Need for Low Faecal Haemoglobin Concentration Estimates
RkJQdWJsaXNoZXIy MTUyODc1Mw==