Amy Carpenter
May 2024—Many laboratories have brought order to chaos in test ordering by launching initiatives to do so, for cost and staff savings and patient care benefits. TriCore is one—it set its sights on orders for monoclonal gammopathies.
“Appropriate testing for monoclonal proteins is very much improved by laboratory-directed testing options,” said TriCore chief scientific officer David G. Grenache, PhD, D(ABCC), in an ADLM annual meeting session last year on method selection and testing in monoclonal gammopathy management.
“When I arrived at TriCore in 2017, our choice of electrophoresis tests was a giant bag of mess,” said Dr. Grenache, medical director of the immunology laboratory and clinical professor of pathology, University of New Mexico School of Medicine. “You could order anything you wanted in any order you wanted.”
Another reason the laboratory set out on this path was to see if it could decrease the time between symptomatology and multiple myeloma diagnosis, Dr. Grenache said. “It’s been shown clearly that a delay in diagnosis leads to more complications.” Kariyawasan, et al., found in a retrospective case review that a prolonged delay before diagnosis is associated with a significant impact on the clinical course (Kariyawasan CC, et al. QJM. 2007;100[10]:635–640). And Koshiaris, et al., in their systematic review and meta-analysis, found that many patients experienced a diagnostic interval longer than three months until diagnosis is confirmed (Koshiaris C, et al. BMJ Open. 2018;8[6]:e019758).
All electrophoresis is performed in TriCore’s core reference laboratory in Albuquerque. “We use capillary electrophoresis for serum proteins, immunotyping, and hemoglobin variants,” Dr. Grenache said. Agarose gel is used for urine protein electrophoresis, urine immunofixation, and hemoglobin when an alternative method is needed to confirm a variant hemoglobin detected by capillary electrophoresis.
In 2017, the orderable test options were serum protein electrophoresis, urine protein electrophoresis (random and 24 hour), immunofixation/immunotyping electrophoresis (urine), and serum free light chains. “And there was no attempt to corral physicians into ordering appropriate test combinations. It was a free-for-all,” Dr. Grenache said.
When they looked at the frequency with which three tests were ordered—protein electrophoresis, immunotyping, and free light chains—they found it to be only three percent, similar to a frequency of 6.1 percent for the same trio of tests found in a survey of laboratories in 2016 (Genzen JR, et al. Arch Pathol Lab Med. 2018;142[4]:507–515).
Still, Dr. Grenache said, “I knew we could make some improvements. Our workflow was labor-intensive because we were not leveraging the technologies and efficiencies.”
Before October 2021, the workflow for serum protein testing in the immunology laboratory began with staff creating a serum protein electrophoresis worklist from the laboratory information system and then testing quality control and patient samples. “That was the easy part,” Dr. Grenache said. The laboratory staff also retrieved folders for any patient who had a history of an abnormal protein and performed preliminary interpretations in the vendor software using a mnemonic built into the LIS. “Then they would have to do a free text of the prior pathologic history. That wasn’t built as a mnemonic because it varies too much.” Additional tasks—electropherogram report printing and matching reports with a pathologic history to the paper folders—and doing the same for immunotyping added up to nine hours each workday, he said. If staff identified a new monoclonal protein in a new patient, “we’d create a new folder. We just kept on creating folders.”
Also found were inconsistencies in interpretation. “We had seven medical directors, with two to three different individuals doing interpretations each week,” he said. “And I like consistency. If I see something on Monday that’s new and I ask for follow-up testing, I want to be the one to see it on Tuesday.” Each medical director spent one to 13 days a month providing interpretations, so some were doing too few to maintain competency.
“We had very little in terms of standardized comments, only what was built in as mnemonics in the LIS,” Dr. Grenache said. Add-on test results weren’t consistently considered by the next interpreter, leading to wasted technologist time and reagents and missed abnormalities.
“Probably worst of all,” he said, “we were doing the interpretations from paper. We had powerful software we could leverage to zoom in on specific areas on those curves, but we weren’t using it.”
Drawing on the advice of others about the importance of laboratory-directed testing guidelines, Dr. Grenache said: “We have to make choosing the right test easier for clinicians. That’s what we started to do.”
He and his colleagues looked to the available guidelines (see summary below). The CAP’s guideline, released online in 2021, came out immediately before TriCore implemented its changes in October 2021 (Keren DF, et al. Arch Pathol Lab Med. 2022;146[5]:575–590). “They were the only ones that said, ‘You should do serum protein electrophoresis and free light chains for the initial detection and then do immunofixation or immunotyping or mass spectrometry to follow up with confirmation,’” Dr. Grenache said. The other organizations’ guidelines recommended tests but no packaging of tests into panels. “It was just, ‘Here are the tests your laboratory should offer,’ but there was no recommendation about, ‘Here’s what should happen and why.’”
After reviewing the available guidelines and considering which workflow would lead to the best patient care, “we got rid of those scattershot tests and created four panels, two for serum and two for urine,” Dr. Grenache said.
IFE, immunofixation electrophoresis; FLC, free light chain; SPE, serum protein electrophoresis; UPE, urine protein electrophoresis; UIFE, urine immunofixation electrophoresis; M protein, monoclonal protein.
Physicians who order monoclonal gammopathy testing in serum can select a monoclonal protein screen, “in which case we do protein electrophoresis, immunotyping, and free light chains,” he said. For patients known to have a history of a monoclonal protein, physicians can order the monitoring panel instead, which is serum protein electrophoresis only. “We will reflexively add immunotyping or immunofixation depending on our interpretation of their electropherogram and history.”
The protein electrophoresis workflow still starts the same way: The worklist is created out of the LIS and the QC and patient samples are tested. “But now we’re leveraging the vendor software for managing these histories.”
Technologists do the preliminary interpretation in the software using statements built into the vendor software. “We no longer have to print reports,” he said, “because we’re not maintaining those histories on paper.” The immunotyping workflow didn’t change much, though they are no longer interpreting from paper and thus don’t have to create folders. “We still have banks of file cabinets because there’s a history there that isn’t always replicated in the software, so we’re still occupying space.”
The time savings? “Three hours in an understaffed laboratory,” Dr. Grenache said. “When you give three hours of time back to your staff, the supervisors and managers are really happy, and the staff are too.”
Instead of seven medical directors, with two to three doing interpretations weekly, there are now four, with one director doing the interpretations each week for a full week.
“We have built standardized interpretive comments into the capillary instrument software,” he said, “and now if on Monday I want to add immunotyping to something, then I see the immunotyping the following day.” He and colleagues wrote standardized interpretive comments that have resulted in more consistent result reporting.
Evaluating electropherogram interpretations in the software has enabled remote interpretation, saving time for medical directors whose offices are not in the core laboratory where testing is done.
Laboratory staff do a preliminary interpretation. To guide their process, Dr. Grenache said, he and colleagues created a protein electrophoresis/immunotyping/immunofixation interpretation algorithm.
Not surprisingly, Dr. Grenache said, “When we started doing this, there was a lot of wailing and gnashing of teeth.”
Clients said, “Why are we doing this anyway? It just adds more confusion. Can we not?” And: “The lab is not going to tell me what to order!” or “If I want immunofixation, I’m going to order it.”
“And we were doing it,” Dr. Grenache said. “They would order immunofixation, we would do immunofixation electrophoresis. But guess what else we had to do? We did protein electrophoresis on all of those samples—even though it wasn’t ordered—so we couldn’t report it, couldn’t bill for it, but we needed it for the interpretation. So we were just doing a lot of free testing.”
Education and communication were a must, of course. A two-sided, one-page document had been sent to all clients explaining “what was going away, what we were building, and why.” But one is just the start. The staff created other explanatory documents for the sales team to leave with clients. “The temperature fell pretty rapidly,” Dr. Grenache said.
They had to provide education, too, on the reporting of the results. “It wasn’t surprising that when we started we would see the screening test ordered on patients who had a history of an abnormal protein, and we would see the monitoring test ordered on patients with no history of a monoclonal protein.”
They put gentle reminders in the interpretive comments to the effect that “Maybe this isn’t the test you intended to order.” After about a year, test ordering improved, but appropriate utilization remains an issue, Dr. Grenache said.
After the October 2021 change, use of the test panel of serum protein electrophoresis, immunotyping, and free light chains increased from a frequency of three percent to 58 percent, he said, which is to be expected when choices are narrowed. “You’ve given people two choices, so of course we’re going to see that increase, but that was the goal. There’s still some variation in how people order, and it’s unavoidable.”
They looked at test use changes by practice—family medicine, nephrology, and oncology. Family medicine’s use of the trio of tests rose from two percent to 61 percent, nephrology from less than one percent to 45 percent, and oncology from 13 percent to 38 percent. The frequency of the use of the monitoring test in oncology increased from 25 percent (serum protein electrophoresis and free light chains) to 37 percent (monoclonal protein monitoring and FLC).
Having bundled the tests in a panel, Dr. Grenache said, there was the expected jump in the number of free light chain and immunotyping tests the laboratory was now performing.
To determine whether the test order changes improved time to multiple myeloma diagnosis, Dr. Grenache and colleagues did a data extraction right before the change—Oct. 13, 2018 through Oct. 12, 2019.
“We had two data sets: pre-change and post-change.” Excluded from these data sets were patients who already had a history of a multiple myeloma or had a bone marrow biopsy before a specified date. “Particularly in the pre-change group, the control group, we grouped test codes together from the same date,” Dr. Grenache said. If serum protein electrophoresis and/or immunofixation electrophoresis were ordered on the same date, they were grouped together in the preset. This didn’t have to be done post-change because they were already grouped.
Definitions were used to identify the number of days from laboratory testing to a diagnosis of multiple myeloma. “First laboratory order” was the earliest order collection date. “First diagnosis date” was identified by the earliest appearance of the ICD-10 code for multiple myeloma. “Days to diagnosis” was calculated by determining the number of days between the first order date and the first diagnosis date.
Before the test order changes (2019–2020), the average number of days to reach a diagnosis of multiple myeloma, based on the appearance of the ICD-10 code, was 133 (n = 82). After the change (2021–2022), the days to diagnosis decreased to 105 (n = 73). “It’s not yet statistically significant,” Dr. Grenache said. “I do think we’re going to get there—we just need more time to increase the number of patients. I’m very optimistic.”
Both immunofixation and immunotyping can detect monoclonal proteins.
David Keren, MD, professor of pathology, University of Michigan, and coauthor of the CAP guideline on laboratory detection and initial diagnosis of monoclonal gammopathies, noted in the Q&A after Dr. Grenache concluded his talk that the question of one technique versus the other is a long-standing one. Dr. Keren cited a Memorial Sloan Kettering Cancer Center study that found immunotyping provides equivalent results to immunofixation in detecting monoclonal proteins, and that, the authors write, “Training and experience are critical to the accurate interpretation of IT” (Thoren KL, et al. J Appl Lab Med. 2021;6[6]:1551–1560).
“In the beta region, immunofixation can work better than immunotyping because when you do immunofixation, you get rid of all the background,” Dr. Keren explained. “There’s no transferrin, no C3 to interfere. But in the gamma region, you don’t have that. When you do immunofixation you have a background of polyclonal. And when you do an immunosubtraction, even if it’s a small M protein, you will be able to see it better in the subtraction because it will be removed in the appropriate one and not in the other. So they’re both good techniques.”
Both IF and IT are used at the University of Michigan, Dr. Keren said. “For the initial screen on serum with no obvious restriction in the electrophoretic pattern, IF is used. We find that the IF result is often ambiguous when evaluating subtle gamma restrictions, and we use IT for those. For beta region bands, IF is used for the initial detection, but IT is helpful in measuring the M protein’s area under the curve” (Keren DF. Electrophoresis and Mass Spectrometry: Monoclonal Gammopathies and Protein Disorders. ASCP Press; 2024).
“No test is perfect,” Dr. Grenache said. “Immunofixation is going to miss some things, and immunotyping is going to miss some things. But when it comes to efficiencies we’ve gained in the laboratory from the workflow, it’s difficult to say I prefer immunofixation.”
At TriCore, the staff were already doing immunotyping and were proficient in it. The vendor provided additional training in interpretation. “It’s important that the people doing the interpretations do them with consistency,” he said. “It’s all about remaining competent.”
Amy Carpenter is CAP TODAY senior editor.