4.28: Tumor Markers

A laboratory performs analyses for prostate specific antigen (PSA) by enzyme immunometric assays. In order to minimize costs, the laboratory performs these analyses twice a week: Tuesday, the day after an oncology clinic and Friday, the day after a urology clinic. On a Wednesday, an irate oncology physician calls the laboratory demanding to know where the result for the PSA test is for his patient. The technologist responsible for this assay looks up the result in the laboratory information system (LIS) and finds that the result are “pending.” The levels of PSA obtained that Tuesday were greater than the highest standard of the assay, and so the samples will be diluted and reanalyzed on Friday. The physician, not especially happy with the delay, agrees to wait until Friday for the results.

On Friday, the technologist in the PSA area performs the analysis on a 1:3 dilution of the sample. The results, once again, exceed the standard curves of the PSA assay and requires still further dilution. The laboratory supervisor, knowing that waiting for the next routine PSA assay would result in an unacceptable delay, asks the technologist to rerun the assays immediately.

QUESTIONS

1. What dilutions should the technologist prepare for the assays this time?
2. The technologist looks up the patient’s previous result PSA. It was 25 times the upper limit of normal. The technologist, who works 4 hours overtime, reports results on the sample that was diluted 1:50. Can you recommend a policy for the analysis of PAP and PSA in this laboratory that will avoid the above situation?

Questions to Consider

1. What analytical factors do you think should be considered when setting up dilutions of samples for subsequent analysis?
2. What clinical factors should be considered when setting up sample dilutions for analysis?

Answer

1. If a technologist is not sure of which dilution to make, then the technologist should set up more than one dilution. Hopefully, this will enable at least one of the dilutions to fit the analytical criteria listed in question number one. If more than one dilution falls within the linear range, a check on the validity of the results can be made by comparing the answers. Probably 1:5, 1:10 and 1:50 dilutions would be advised for this sample. This sample was drawn from a patient attending an oncology (cancer) clinic so one might predict very elevated levels of these tumor markers.
2. The laboratory should actively use the capability of most laboratory information system (LIS, Chapter 18) to review the previous results for tests that may have very high and that are infrequently performed. Knowing the previous result of a patient who attends a disease treatment clinic can provide information that can be used when choosing dilutions for a current specimen. If an LIS does not have this capability, a manual “delta” check file on index cards can be uesed.
   
   To minimize delays when samples are drawn from patients attending a clinic for the first time, the laboratory’s policy might require that the technologists perform the PSA analysis on an undiluted sample, a sample diluted two times, and a sample diluted ten times. Although this may appear to consume unnecessary resources, it will reduce turnaround time, repeat analyses, and perhaps the need for overtime laboratory work.

Note

More modern immunometric assays for PSA have larger reportable ranges and many modern instruments have the capability to automatically re-run the analysis for samples whose results are greater than the upper limit of the reportable range. The second analysis can be performed on an actual dilution or simply on a smaller volume of sample.

Answers to Questions to Consider

1. The important analytical factors to consider when preparing dilutions include:
   • Linearity of the analytical reaction: The aim is to dilute the analyte to a level within the established linear range, also called the reportable range, of the assay. Remember, all assays have lower limits of linearity; one never reports out '0' for a result, but "less than" (<). This lower limit is sometimes referred to as the lower limit of sensitivity of the assay.
   • Analytical sensitivity: One does not want to over-dilute the analyte to a level below the limit of sensitivity; the final result is often bizarre if this caution is not observed.
   • Precision: One usually gets better precision with a larger signal (e.g., absorbance, Chapter 23; or counts per minute for radioactivity, p 198-Chapter 9). Therefore, one would not want to dilute the analyte down into the very lowest region of the reportable range. In addition, each pipetting step for dilution has some inherent imprecision which is now added onto the assay’s degree of imprecision. Larger error assocaited with the new analysis can affect the final result, sometimes producing erroneous results. Thus the dilution should be one that is easily attainable with the resources at hand to produce an accurate dilution.
   • Sample volume: One needs sufficient volume of sample to produce the optimal dilution. If insufficient sample is available, a larger, less optimal dilution may be necessary.
   • The “hook” effect: Many immunoassays have an anomolous “hook” effect which occurs at very high levels of analyte and results in lower results than the true value. A “hook” effect is another limit of linearity. One wants to dilute an analyte measured by an immunoassay below the region of a “hook” effect.
2. Important clinical factors to be considered when preparing dilutions include:
   • The levels of analyte that are usually associated with specific disease states.
   • The “usual” results obtained on a specific population or an individual.
   • The current level of that analyte in the specific patient in question.