Best LIS for Molecular and PCR Labs

Part 493 has one line written for amplification assays, inside the control procedures standard. Unless the lab runs an approved alternative (in practice an IQCP under Appendix C of the State Operations Manual), it applies “at least once each day patient specimens are assayed or examined” § 493.1256(d)(3):

An assay with an extraction step also needs “two control materials, including one that is capable of detecting errors in the extraction process” § 493.1256(d)(3)(iv). If the assay's instructions ask for controls on every plate, that is more than the daily minimum, and CLIA holds the lab to the manufacturer's number and frequency when they meet or exceed the rule § 493.1256(d)(2). Control results must be acceptable “before reporting patient test results” § 493.1256(f), and “The laboratory must document all control procedures performed” § 493.1256(g).

That is where the plate map earns its keep. If the LIS knows which wells hold which controls, it can tie every patient well to the controls that ran beside it, which a spreadsheet stuck to the cycler cannot.

Of the six vendors, only Dendi documents the whole chain step by step in a public help centre. Its plate-map screen asks for the controls first, well by well, then the samples, scanned or pasted from a liquid-handler file (Plate Maps). Its instrument settings can require a verified QC run before patient data is imported (Instruments), and a verified run records the run time, the verification time and the user who verified it (Control Sets). CrelioHealth's help centre shows each device's QC values held for review before they reach the Levey-Jennings chart. Prolis's QC feature page says “QC must pass before patient results can be released” and lists timestamps, lot numbers and acceptance criteria for every control run. For QC software that sits beside an LIS, see the lab QC software comparison.

The room matters too. Amplification “not contained in closed systems” must follow a “uni-directional workflow” with separate areas for specimen preparation, amplification and detection § 493.1101(a)(3). No LIS makes a lab one-way, but batch and plate timestamps are the record of where a run was and when.

A real-time PCR instrument hands the LIS a cycle threshold (Ct) for each target in each well. The physician gets a word: detected, not detected, or a reason the well could not be read. Somewhere between the two, a rule turns a number into that word, and that rule is a piece of the test system.

CLIA asks the report to show “The test result and, if applicable, the units of measurement or interpretation, or both” § 493.1291(c)(6). It does not require the Ct itself. If you print it, the lab must also be ready, on request, to give “information that may affect the interpretation of test results” § 493.1291(e), which is a reason to put a line on the report saying what a Ct is and is not.

Those specifications are where your cutoff comes from. A lab running the manufacturer's cutoff verifies it; a lab that changes it, or built the assay, establishes its own before reporting § 493.1253(b)(2). Either way the number typed into the LIS test build has to match the validation file, and a change to it is a change to the test.

Dendi's help centre shows how one LIS stores that rule. A test is set as cutoff-based, and when “Positive/Detected above cutoff” is left unchecked, “a quantitative result value below the set cutoff will indicate a positive/detected result”: the Ct logic. A multi-target assay is resolved by a truth table that sets the final interpretation for every combination of target results, and a cutoff can be set per instrument (Lab Menu: Tests, Instruments). None of the other five vendors shows Ct handling in a document we could read; Prolis lists “result interpretation protocols” for molecular labs.

The definition puts the rule in the lab's standard operating procedure, so the LIS rule should read the same as the SOP. In a PCR lab that might be a respiratory panel added after a negative rapid screen, or a genotyping assay added after a detected result. Dendi builds a reflex from a trigger (a result value) and a reflex test profile added to the order; LigoLab routes work and triggers reflex tests by laboratory-defined rules; NucleoLIS places reflex orders from rules on the specimen, the referring provider or the case type. Whatever the system, the requisition still has to name the tests ordered § 493.1241(c)(4), so check that the reflex appears on the order and the report.

Pooled testing

Pooling mixes several specimens into one reaction and retests the members only when the pool is not negative. It saves reagent when positives are rare. If the assay's instructions for use do not cover pooling, pooling changes the test system, and the lab must establish its own performance specifications before reporting § 493.1253(b)(2).

The LIS then has to keep positive identification of every specimen § 493.1283(a)(1) while the plate only sees the pool. Dendi is the one vendor here that documents it: a bulk pooled order prints labels for the pool and for each member order, and by default a pool resulted Not Detected copies that result to every member and verifies it, while any other result moves the members back to “Pending Results” with a comment that each needs testing (Pooled Orders). Dendi sells pooling as a premium feature (Pooled Testing).

“Records of patient testing including, if applicable, instrument printouts, must be retained” § 493.1283(b). For a PCR lab the printout is the run file: the plate setup, the amplification curves and the Ct table the cycler exported. Ask whether the LIS stores that file with the batch or only the results it parsed from it; if only the results, the lab has to archive the run files somewhere else for two years.

Test procedures stay for 2 years after a procedure is discontinued § 493.1105(a)(2), so an assay retired this year still needs its SOP, and its cutoff, on file. A molecular result that goes out inside a pathology report sits on a document CLIA asks the lab to keep for 10 years § 493.1105(a)(6)(ii); the pathology LIS comparison covers that clock.

Vendor demos run on clean sample data. Bring your own: a real run file, your control layout, your cutoffs. Each step below tests one row of the plate map against the rule behind it.

The same script works for a general LIS that claims molecular support; compare it with the CLIA rows in the LIS comparison for small clinical labs.