PC-Alarm Focus scientific background
Find out about the measurement method and the available study data. You can also show this page to your doctor when discussing the place of PC-Alarm Focus in your investigation.
PC-Alarm Focus looks for the signal of one specific variant of the TMPRSS2–ERG gene fusion in a urine sample. The method is developed in-house by European Life Technologies Hungary.
The development data set contains the records of 178 patients. The detailed analysis covered 86 patients; 75 of them had at least one evaluable tissue comparison.
The figures below are positive and negative percent agreement with the given tissue method – not clinical sensitivity or specificity – and come from a retrospective, exploratory analysis.
The measurement targets the T1-E4 transcript of the TMPRSS2–ERG gene fusion. RNA preparation is followed by a targeted, two-step PCR-based measurement. In today's routine we measure two internal controls in every analytical run, to check whether the RNA can be analysed and that the sample is of prostate origin. This description of the routine does not mean that both controls are documented in every record of the development data set. The method examines the signal of one specific fusion variant; it does not measure inherited predisposition and does not give a numerical cancer risk.
In the detailed comparison, an evaluable immunohistochemical comparison was available in 44 cases, tissue qPCR in 74 cases and at least one of the two in 75 cases; 11 patients had no evaluable tissue reference. A substantial part of the group studied consisted of patients already diagnosed and awaiting surgery. This group composition is the frame for interpreting the results: the proportions below apply to this population, not to the general public. Data collection in the pre-biopsy population is ongoing.
Hungarian clinical partners took part in collecting the development data set: Semmelweis Hospital in Kiskunhalas, Péterfy Sándor Street Hospital and the University of Szeged.
Of the 172 patients with a urine result in the development data set, 168 results could be classified according to the analysis rule (97.7%; 95% CI 94.2–99.1%; retrospective, exploratory data). This is result classifiability, not a validated analytical sample-adequacy or routine success rate.
We compared the urine result with ERG immunohistochemistry and tissue fusion qPCR. The analysis shows positive and negative agreement with the results of the tissue samples examined. The table comes from a retrospective, exploratory analysis; the confidence intervals indicate the uncertainty of the estimates. The detailed result rule and the limits of interpretation can be read below the table.
The development data set contains samples from 178 real patients. The detailed comparative analysis was made on the data of 86 of them; 75 of those had at least one evaluable tissue comparison available, 11 did not. The number of patients therefore differs from one comparison row to the next – the table states the denominator for every row. The breakdown by cancer status covers 82 of the 86 patients. The comparison group that also includes the negative controls is a separate, broader group.
| Tissue comparison | Evaluable pairs | Positive agreement: urine-positive among tissue-positive patients | Negative agreement: urine-negative among tissue-negative patients |
|---|---|---|---|
| ERG immunohistochemistry (IHC) | 44 | 4/6 = 66.7%; 95% CI 30.0–90.3% | 37/38 = 97.4%; 95% CI 86.5–99.5% |
| Tissue FFPE fusion qPCR | 74 | 8/15 = 53.3%; 95% CI 30.1–75.2% | 55/59 = 93.2%; 95% CI 83.8–97.3% |
| Composite according to the available tissue results | 75 | 8/15 = 53.3%; 95% CI 30.1–75.2% | 56/60 = 93.3%; 95% CI 84.1–97.4% |
The table shows how the urine measurement agreed with the available tissue results. The individual methods do not describe the fusion status of the whole prostate; the agreement rates are therefore not the same as the diagnostic sensitivity or specificity for prostate cancer.
How results were formed in the analysis: we used the main urine result cell and treated any positive marking as positive. The comment field alone did not override this classification; even so, the uncertain or conflicting cases cannot all be treated as closed routine findings. The composite tissue result is positive if any available tissue test is positive, and negative if the available results are negative. This is why the sample-pair counts differ across the table.
Of the four urine-positive, tissue-negative cases, three had histologically confirmed cancer. The fourth patient was being treated for benign prostatic enlargement; the resection tissue sample was negative by both tissue methods, the repeat urine measurement gave an equivocal result, and clinical follow-up is ongoing. The reason for tissue negativity – for example the focal distribution of a tumour, low expression or the site of sampling – cannot be established from this data set; it is a hypothesis.
The tissue comparison has its own limits: IHC marks a protein, tissue qPCR measures one specific fusion transcript, and the tissue sample taken represents only part of the prostate. A negative tissue result is therefore not proof that the whole prostate is free of the fusion. This is a methodological limit, not evidence that the differing urine result is the correct one.
A positive urine result increases the odds of the tissue fusion being present by a factor of 8–25 (positive likelihood ratio, LR+ 8–25 across the three comparisons). This applies to the tissue fusion, not to cancer risk, and cannot be transferred to men with a raised PSA.
IHC marks the ERG protein, while tissue qPCR looks for the targeted fusion transcript. The groups overlap; the composite uses the same tissue results. The three rows are therefore not three independent validations. Among patients with both tissue tests (n=43), positive agreement is 4/8 = 50.0% (95% CI 21.5–78.5%) and negative agreement is 34/35 = 97.1% (95% CI 85.5–99.5%). These too are agreement measures against the tissue samples examined, from an exploratory analysis.
Among the 64 patients classified as having cancer in the group studied, the urine test gave a positive signal in 13 (20.3%; 95% CI 12.3–31.7%). This is the positivity rate observed in the group, not sensitivity for the target population. The breakdown by cancer status covers 82 of the 86 patients: 64 classified as having cancer and 18 as not.
The 13/64 ratio is the urine positivity observed in the patient group examined, not a general cancer detection rate. Not every tumour carries the targeted gene fusion. Tissue qPCR and the urine measurement target the same T1-E4 site, so they may share variant and measurement limitations; IHC is a different comparator, at protein level. The effects of the composition of the group, of the fusion variant and of detectability in urine cannot be reliably separated from this data set.
We examined the urine-positive, tissue-negative cases separately. The numbers for positive and negative agreement are in the table above, and the interpretation of the discordant cases is in the explanation that goes with it.
In a separate, broader comparison group – patients with benign prostatic enlargement together with all evaluable negative controls – urine positivity was 4/35 (11.4%; 95% CI 4.5–26.0%). This group is not the same as the 18-patient subgroup above, and the negative controls had no histological verification, so we do not derive a standalone clinical specificity claim from it.
A negative result does not rule out cancer: not every tumour carries the targeted alteration, and even when the signal is present it is not always detectable in urine.
PSA is the blood level of a protein characteristic of the prostate. It can rise because of cancer, but also because of benign prostatic enlargement, inflammation or other, non-cancerous causes. PC-Alarm Focus, by contrast, looks in urine for the imprint of a somatic genetic alteration, the TMPRSS2–ERG gene fusion.
The TMPRSS2–ERG gene fusion is a molecular alteration characteristic of prostate tumours; it can also occur in HGPIN lesions. In the tissue samples of our own BPH subgroup we found 0/14 positive by ERG IHC and 0/14 positive by tissue fusion qPCR (95% CI 0–21.5% for both; exploratory analysis). These are data on the tissue samples examined. They do not prove that PC-Alarm Focus detects cancer more accurately than an elevated PSA.
A negative result does not rule out prostate cancer. The test complements PSA and urological examination; it does not replace them. The result must be interpreted together with the clinical situation.
In the PCPT study, among its defined participants aged 62–91 with a PSA of at most 4 ng/ml, cancer was found by biopsy in 15.2%. This is not a prevalence figure for today's general interested population, nor a performance study of PC-Alarm Focus. What it does show is that a PSA value on its own does not rule out the presence of cancer – one of the reasons why a measurement target different from PSA can provide information of its own.
The table below estimates how the urine test result modifies the probability of prostate cancer within a given PSA band. It is a model calculation, not a measurement: its inputs stand separately and their uncertainties add up. A consumer reading of the same table appears on the test page.
| PSA (ng/ml) | Without a PC-Alarm Focus test | Negative PC-Alarm Focus test | Positive PC-Alarm Focus test |
|---|---|---|---|
| 0 – 0,5 | 7 | 6 | 9 |
| 0,6 – 1,0 | 10 | 10 | 14 |
| 1,1 – 2,0 | 17 | 16 | 23 |
| 2,1 – 3,0 | 24 | 23 | 31 |
| 3,1 – 4,0 | 27 | 26 | 34 |
| 4 – 10 | 30 | 27 | 55 |
| 10 – 20 | 65 | 61 | 84 |
The inputs: as sensitivity, the 20.3% urine positivity observed in the patient group examined; as specificity, 85.7% below and 92.9% above a PSA of 4 ng/ml. No bias correction is applied; this is the conservative variant.
The baseline frequencies: the bands below 4 ng/ml from Thompson et al., N Engl J Med 2004;350:2239–2246 (2,950 participants, biopsy at the end of the trial); the 4–10 band from Presti JC, AUA 2006, abstract 468 (midpoint of the 25–45% range); the 10–20 band from Catalona WJ et al., N Engl J Med 1991;324:1156–1161. The band above 20 ng/ml rested on an estimate with no primary source behind it, and is therefore absent from the table.
“Cancer” here means every histologically confirmed prostate cancer, including the slow-growing forms that need no treatment. The table does not take account of age, family history, the digital rectal examination or imaging findings.
Other urine-based tests, for example ExoDx, SelectMDx and MyProstateScore, estimate the risk of clinically significant prostate cancer using different biomarkers in defined patient groups, and can support the decision about biopsy. PC-Alarm Focus looks for the signal of one specific gene fusion. Because the endpoints differ, the performance figures cannot be compared directly; no comparative study of our own has been carried out.
The FDA statistical guidance on diagnostic tests also distinguishes the roles of a reference standard and a comparator method; we use it as a methodological source. The European Association of Urology (EAU) evaluates tests in the context of the clinical situation and the specific method.
What these sources are for: the FDA guidance provides the methodological framework for interpreting agreement measures, and the EAU guideline describes the clinical setting of prostate investigation. The Thompson paper is background on the limits of PSA. None of them is a product authorisation or an independent clinical validation of PC-Alarm Focus.
Publications on the biology of the marker – none of them a performance study of PC-Alarm Focus:
With a positive result a urological consultation is recommended; the result is not a diagnosis. A negative result does not replace a warranted work-up.
