MatchMiner: an open-source platform for cancer precision medicine

Widespread, comprehensive sequencing of patient tumors has facilitated the usage of precision medicine (PM) drugs to target specific genomic alterations. Therapeutic clinical trials are necessary to test new PM drugs to advance precision medicine, however, the abundance of patient sequencing data coupled with complex clinical trial eligibility has made it challenging to match patients to PM trials. To facilitate enrollment onto PM trials, we developed MatchMiner, an open-source platform to computationally match genomically profiled cancer patients to PM trials. Here, we describe MatchMiner’s capabilities, outline its deployment at Dana-Farber Cancer Institute (DFCI), and characterize its impact on PM trial enrollment. MatchMiner’s primary goals are to facilitate PM trial options for all patients and accelerate trial enrollment onto PM trials. MatchMiner can help clinicians find trial options for an individual patient or provide trial teams with candidate patients matching their trial’s eligibility criteria. From March 2016 through March 2021, we curated 354 PM trials containing a broad range of genomic and clinical eligibility criteria and MatchMiner facilitated 166 trial consents (MatchMiner consents, MMC) for 159 patients. To quantify MatchMiner’s impact on trial consent, we measured time from genomic sequencing report date to trial consent date for the 166 MMC compared to trial consents not facilitated by MatchMiner (non-MMC). We found MMC consented to trials 55 days (22%) earlier than non-MMC. MatchMiner has enabled our clinicians to match patients to PM trials and accelerated the trial enrollment process.

Abstract 3382: MatchMiner: An open-source computational platform for genomically-driven matching of cancer patients to precision medicine clinical trials

To facilitate interpretation of complex tumor sequencing data and clinical trial genomic eligibility criteria, we developed MatchMiner, an open-source platform to computationally match cancer patients to precision medicine clinical trials based on clinical and genomic features. MatchMiner supports two distinct workflows: (1) patient-centric mode, in which an oncologist can find clinical trial matches for a specific patient, and (2) trial-centric mode, in which a clinical trial investigator can identify and recruit patients for a specific trial.

MatchMiner has been operational at Dana-Farber Cancer Institute since early 2017. There are currently 275+ trials curated in the system and genomic data from 26,000+ patients. Over 80% of living patients match to at least one open clinical trial, with an average of 6 trial matches per patient. At least 98 patients have enrolled on a clinical trial as a result of MatchMiner.

To enable computational matching, we developed clinical trial markup language (CTML), a structured format to encode detailed information about a trial. CTML utilizes boolean logic to define clinical (e.g. cancer type), demographic (e.g. age) and genomic (e.g. specific mutations, copy number alterations, structural variants or mutational signatures) eligibility, which can be applied to individual arms of a trial.

MatchMiner is an open-source two-tier web application with a Python-based REST API server and an AngularJS front-end. MatchMiner utilizes Security Assertion Markup Language (SAML)-based authentication and is fully HIPAA-compliant when hosted behind a secure institutional firewall. MatchMiner ingests clinical and genomic data, and connects to existing clinical systems, including clinical trial management systems for real-time trial status.

We recently refactored the core matching algorithm (the matchengine), which improves upon the original matchengine in several ways: (1) increased granularity in reporting the reason for a match; (2) can match all patients/trials or individual patients/trials; (3) easily extensible to match based on additional data types.

The MatchMiner open-source software package is available through GitHub (https://github.com/dfci/matchminer). We are committed to supporting MatchMiner as an open-source software; to our knowledge, at least five cancer centers are implementing MatchMiner at their own institutions.

In summary, we have defined a standard for encoding clinical trial information in a structured and computable form, and we have developed an open-source computational trial matching platform to support patient-specific trial identification as well as trial-specific patient recruitment. We are actively collaborating with clinical groups at Dana-Farber Cancer Institute and other institutions to understand the role of MatchMiner in their clinical workflows, and we are committed to continuing to evolve MatchMiner to meet clinical needs.

Citation Format: Tali Mazor, Rachel B. Keller, Priti Kumai, James Lindsay, Eric Marriott, Andrea Ovalle, Ethan Siegel, Joyce Yu, Michael Hassett, Ethan Cerami. MatchMiner: An open-source computational platform for genomically-driven matching of cancer patients to precision medicine clinical trials [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3382.

Abstract A024: MatchMiner: An open-source computational platform for genomically-driven matching of cancer patients to precision medicine clinical trials

To facilitate interpretation of complex tumor sequencing data and clinical trial genomic eligibility criteria, we developed MatchMiner, an open-source platform to computationally match cancer patients to precision medicine clinical trials based on clinical and genomic features. MatchMiner supports two distinct workflows: (1) a patient-centric mode, in which an oncologist can find clinical trial matches for a specific patient, and (2) a trial-centric mode, in which a clinical trial investigator can identify and recruit patients for a specific trial. MatchMiner has been operational at Dana-Farber Cancer Institute since early 2017. There are currently 250+ trials curated in the system and genomic data from 24,000+ patients. Over 82% of living patients match to at least one open clinical trial, with an average of 6 trial matches per patient. At least 85 patients have enrolled on a clinical trial as a result of MatchMiner. The MatchMiner open-source software package is available through GitHub (https://github.com/dfci/matchminer). MatchMiner is a two-tier web application with a Python-based REST application programming interface (API) server and an AngularJS front-end. MatchMiner utilizes Security Assertion Markup Language (SAML)-based authentication and, when hosted behind a secure institutional firewall, is fully HIPAA-compliant. MatchMiner can connect to existing clinical systems, including clinical trial management systems for real-time trial status. To enable computational matching to clinical trials, we developed clinical trial markup language (CTML), a structured format to encode detailed information about a trial. CTML utilizes boolean logic to define clinical (e.g. cancer type), demographic (e.g. age) and genomic (e.g. specific mutations, copy number alterations, structural variants or mutational signatures) eligibility, which can be applied to individual arms of a trial. We recently refactored the core matching algorithm (the matchengine), which improves upon the original matchengine in several ways. While the original matchengine reported the reason for a patient-trial match, the refactored matchengine provides additional, more granular details. In addition, the original matchengine ran at the cohort level, matching all patients to all trials, whereas the refactored matchengine can also run against individual patients or trials, speeding up the matching process. The refactored matchengine is also easily extensible to match based on additional data types. In summary, we have defined a standard for encoding clinical trial information in a structured and computable form, and we have developed an open-source computational trial matching platform to support patient-specific trial identification as well as trial-specific patient recruitment. We are actively collaborating with clinical groups at Dana-Farber Cancer Institute to understand the role of MatchMiner in their clinical workflows, and we are committed to continuing to evolve MatchMiner to meet clinical needs.

Citation Format: Tali Mazor, Rachel B Keller, Priti Kumari, James Lindsay, Eric Marriott, Andrea Ovalle, Ethan Siegel, Elizabeth H Williams, Joyce Yu, Michael Hassett, Ethan Cerami. MatchMiner: An open-source computational platform for genomically-driven matching of cancer patients to precision medicine clinical trials [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference on Molecular Targets and Cancer Therapeutics; 2019 Oct 26-30; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2019;18(12 Suppl):Abstract nr A024. doi:10.1158/1535-7163.TARG-19-A024

Lateral trunk lean gait modification increases the energy cost of treadmill walking in those with knee osteoarthritis

OBJECTIVE

To compare the energy expenditure of increased lateral trunk lean walking - a suggested method of reducing medial compartment knee joint load - compared to normal walking in a population of older adults with medial knee osteoarthritis (OA).

METHOD

Participants completed two randomly-presented treadmill walking conditions: 15 min of normal walking or walking with ten degrees of peak lateral trunk lean. Lateral trunk lean angle was displayed in front of the participant in real-time during treadmill conditions. Energy expenditure (VO2 and METs), heart rate (HR), peak lateral trunk lean angle, knee pain and perceived exertion were measured and differences between conditions were compared using paired t-tests.

RESULTS

Twelve participants (five males, mean (standard deviation (SD)) age 64.1 (9.4) years, body mass index (BMI) 28.3 (4.9) kg/m²) participated. All measures were significantly elevated in the lateral trunk lean condition (P < 0.008), except for knee pain (P = 0.22). Oxygen consumption (VO2) was, on average 9.5% (95% CI 4.2-14.7%) higher, and HR was on average 5.3 beats per minute (95% CI 1.7-9.0 bpm) higher during increased lateral trunk lean walking.

CONCLUSION

Increased lateral trunk lean walking on a treadmill resulted in significantly higher levels of steady-state energy expenditure, HR, and perceived exertion, but no difference in knee pain. While increased lateral trunk lean has been shown to reduce biomechanical measures of joint loading relevant to OA progression, it should be prescribed with caution given the potential increase in energy expenditure experienced when it is employed.

Quantifying lower limb joint position sense using a robotic exoskeleton: a pilot study

Clinicians and scientists often focus on tracking the recovery of motor skills after spinal cord injury (SCI), but less attention is paid to the recovery of sensory skills. Measures of sensory function are imperative for evaluating the efficacy of treatments and therapies. Proprioception is one sensory modality that provides information about static position and movement sense. Because of its critical contribution to motor control, proprioception should be measured during the course of recovery after neurological injury. Current clinical methods to test proprioception are limited to crude, manual tests of movement and position sense. The purpose of this study was to develop a quantitative assessment tool to measure joint position sense in the legs. We used the Lokomat, a robotic exoskeleton, and custom software to assess joint position sense in the hip and knee in 9 able-bodied (AB) subjects and 1 person with incomplete SCI. We used two different test paradigms. Both required the subject to move the leg to a target angle, but the presentation of the target was either a remembered or visual target angle. We found that AB subjects had more accurate position sense in the remembered task than in the visual task, and that they tended to have greater accuracy at the hip than at the knee. Position sense of the subject with SCI was comparable to those of the AB subjects. We show that using the Lokomat to assess joint position sense may be an effective clinical measurement tool.

'Sink or swim': an evaluation of the clinical characteristics of individuals with high bone mass

SUMMARY

High bone mineral density on routine dual energy X-ray absorptiometry (DXA) may indicate an underlying skeletal dysplasia. Two hundred fifty-eight individuals with unexplained high bone mass (HBM), 236 relatives (41% with HBM) and 58 spouses were studied. Cases could not float, had mandible enlargement, extra bone, broad frames, larger shoe sizes and increased body mass index (BMI). HBM cases may harbour an underlying genetic disorder.

INTRODUCTION

High bone mineral density is a sporadic incidental finding on routine DXA scanning of apparently asymptomatic individuals. Such individuals may have an underlying skeletal dysplasia, as seen in LRP5 mutations. We aimed to characterize unexplained HBM and determine the potential for an underlying skeletal dysplasia.

METHODS

Two hundred fifty-eight individuals with unexplained HBM (defined as L1 Z-score ≥ +3.2 plus total hip Z-score ≥ +1.2, or total hip Z-score ≥ +3.2) were recruited from 15 UK centres, by screening 335,115 DXA scans. Unexplained HBM affected 0.181% of DXA scans. Next 236 relatives were recruited of whom 94 (41%) had HBM (defined as L1 Z-score + total hip Z-score ≥ +3.2). Fifty-eight spouses were also recruited together with the unaffected relatives as controls. Phenotypes of cases and controls, obtained from clinical assessment, were compared using random-effects linear and logistic regression models, clustered by family, adjusted for confounders, including age and sex.

RESULTS

Individuals with unexplained HBM had an excess of sinking when swimming (7.11 [3.65, 13.84], p < 0.001; adjusted odds ratio with 95% confidence interval shown), mandible enlargement (4.16 [2.34, 7.39], p < 0.001), extra bone at tendon/ligament insertions (2.07 [1.13, 3.78], p = 0.018) and broad frame (3.55 [2.12, 5.95], p < 0.001). HBM cases also had a larger shoe size (mean difference 0.4 [0.1, 0.7] UK sizes, p = 0.009) and increased BMI (mean difference 2.2 [1.3, 3.1] kg/m(2), p < 0.001).

CONCLUSION

Individuals with unexplained HBM have an excess of clinical characteristics associated with skeletal dysplasia and their relatives are commonly affected, suggesting many may harbour an underlying genetic disorder affecting bone mass.

Concentrated ascitic fluid reinfusion in cirrhotic patients: a simplified method

A new method for ascites filtration and reinfusion, which uses a single Cuprophan filter and is performed in the dialysis unit, is reported. Thirty-one procedures were performed in 17 patients with cirrhosis and massive ascites. A mean volume of 8.6 L of ascitic fluid was removed; from this volume, 5 L were ultrafiltered and a concentrated ascitic fluid was reinfused (x = 359.8 mL). The whole procedure was completed in a mean time of 248 minutes. No relevant method-related complications were detected. Moreover, no significant changes in blood urea nitrogen (BUN), creatinine, plasma and urinary electrolytes, or platelet count were found, even in the case of repeated procedures (two to nine times). The reinfused fluid contained a mean value of albumin of 4.7 g/dL and significant amounts of globulins and complement. The overall cost of the materials used in the procedure ($49.46) offered competitive advantages with respect to other types of frequently used methods. In conclusion, we present a safe, effective, and time- and cost-saving technique for ascites reinfusion that represents an advantageous alternative to more complicated and expensive methods or to the currently used medical therapy.

Undiagnosed hepatitis C virus infection in hemodialysis patients: value of HCV RNA and liver enzyme levels

At present, routine screening for hepatitis C virus (HCV) infection is based on the detection of antiviral antibodies. Underdiagnosis of HCV infection by using HCV antibody tests, however, still occurs. Additional diagnostic means are provided by the polymerase chain reaction (PCR). The measurement of aminotransferase (ASAT and ALAT) has served as an auxiliary, less specific test. The present research aimed to design practical and low cost strategies to diminish underdiagnosis of HCV infection in dialysis patients. With this purpose in mind, we examined whether aminotransferases values in HCV antibody-negative patients could be related to undiagnosed HCV infection, by using HCV RNA testing by PCR as the gold standard. In 112 hemodialysis patients, we found 78 negative and 34 positive for HCV antibodies. A major finding was that 222 (28.2%) out of the 78 HCV antibodies-negative patients had positive HCV RNA by PCR. In repeated samples taken at six months follow-up from 19 out of these 22 patients, only one of them was positive for anti-HCV antibodies; moreover, a positive HCV RNA by PCR was confirmed in 13 (68.5%) of them. Within the HCV antibody-negative group, the mean values of ASAT, ALAT and gammaglutamiltransferase were higher (P < 0.001, P < 0.001 and P < 0.02, respectively) in the HCV PCR-positive versus the HCV PCR-negative patients. No significant differences were found in the liver enzyme values between the HCV antibody-negative, HCV RNA positive and the HCV antibody positive, HCV RNA positive individuals. Histological samples from two HCV RNA positive, HCV antibody-negative patients disclosed the presence of a mild liver disease. In conclusion, the present study demonstrates the critical importance of HCV RNA determination by PCR in hemodialysis patients who have no detectable circulating antibodies against the HCV. Furthermore, in conditions in which PCR technology is not readily available, we have established that the existence of a moderate increase of aminotransferases is a helpful clue to detect patients with absent HCV antibodies, and might represent an useful, low cost tool for HCV screening in dialysis patients.

Risks and benefits of interferon-alpha in the treatment of hepatitis

The high worldwide prevalence of chronic viral hepatitis, as well as its progressive course, have led to the performance of multiple clinical studies. The natural history of the disease is different depending on the infecting virus; thus, the evolution to liver cirrhosis and/or hepatocellular carcinoma for the hepatitis B, C and delta (D) viruses in chronic hepatitis is 15, 20 and 75%, respectively. Different therapeutic agents have been used in attempts to modify the natural course of these diseases, interferon-alpha (IFN alpha) having proved to be the most effective. In 30 to 50% of patients, treatment with IFN alpha has achieved inhibition of viral replication, as well as normalisation of aminotransferase levels. Moreover, in a majority of patients, histological improvement is observed, principally in piece-meal necrosis and portal inflammation. The dosage currently recommended for treatment of chronic hepatitis B is 30 to 35MU weekly for a minimum of 4 months; when there is a co-existing delta virus infection, the total dosage employed should be greater. For hepatitis C, the minimum effective dosage is 9MU weekly, and a treatment duration of 12 months is recommended. The administration of IFN alpha produces a series of dose-dependent adverse effects, which are reversible on suspension of the medication. The most frequent of these adverse reactions is the 'flu-like' syndrome, which is self-limited and generally well tolerated. Secondary haematological alterations (leucopenia and thrombocytopenia) are the most frequent cause of reduction in dosage or suspension of treatment, although the latter is not normally necessary. The proportion of patients requiring dosage modification or suspension of treatment fluctuates between 5 and 15%. Taking the evolution of chronic hepatitis into account, there can be no doubt that all patients with this disease should be offered treatment. At present, the drug of choice is IFN alpha, as it slows disease progression and it is generally well tolerated.

Hepatitis B and D genomes in hepatitis B surface antigen negative patients with chronic hepatitis C

Hepatitis B and hepatitis D viral genomes were tested by nested polymerase chain reaction in the serum and liver of 69 hepatitis B surface antigen (HBsAg) negative, anti-hepatitis C virus (HCV) positive patients (47 with HCV RNA and 22 without HCV RNA). Serum hepatitis B virus (HBV) DNA-was detected in 49% of the patients with HCV-RNA and in 64% of those without HCV-RNA. Furthermore, intrahepatic HBV-DNA was found in four of five (80%) of the biopsies analysed. Delta genome was found in 72% and 73%, respectively, of the anti-HCV positive patients with or without HCV-RNA. In addition, intrahepatic delta virus genome was detected in another four liver biopsies studied. In the group of patients with HCV-RNA, the simultaneous presence of hepatitis B and D genomes was statistically higher in transfused patients than in drug addicts, or in those with an unknown infection route (P < 0.001). These results show a high percentage of B and D genomes in HBsAg negative patients with anti-HCV, irrespective of the presence or absence of the HCV genome. However, the clinical implications of this finding should be examined in future studies.

Porphyria cutanea tarda and hepatitis C and B viruses infection: a retrospective study

Based on the knowledge that patients with porphyria cutanea tarda (PCT) usually have chronic liver disease, several authors studied a possible relationship to hepatotropic virus infections. However, the prevalence of hepatitis B virus (HBV)-DNA by polymerase chain reaction (PCR) in serum of these patients, as well as the presence of hepatitis C virus (HCV)-RNA in paired liver, peripheral blood mononuclear cells (PBMCs), and serum samples in these patients has not been reported. We have studied 34 patients with sporadic PCT. Antibodies against HBV were detected in 91% of the patients, but in only 41% of the patients against HBV (P < .01). Viral genomes of HCV and HBV were detected in 65% and 40% of our patients, respectively (P < .05). Genomic and antigenomic HCV strands were found in liver biopsy specimens (100% and 54%), mononuclear cells (100% and 54%), and serum (45% and 0%) from 11 patients. Twelve patients were retrospectively studied, and no correlation was observed between the appearance or disappearance of viral genomes and the simultaneous presence of both genomes with the course of porphyria. In our patients with PCT, detection of viral genomes did not correlate with phlebotomy or length of time since PCT was diagnosed. Our findings demonstrate that HCV infection may be underestimated when detection is performed only in serum of PCT patients, and that HBV infection might also be increased in PCT.

Positive and negative hepatitis C virus RNA strands in serum, liver and peripheral blood mononuclear cells in anti-HCV patients: relation with the liver lesion

To investigate the replicative hepatitis C virus status and its relation to liver damage, serum, peripheral blood mononuclear cells and liver-paired samples from 45 untreated hepatitis C virus infected patients (38 with chronic hepatitis, three with minimal changes, and four with normal liver) were studied by nested polymerase chain reaction, using primers from the 5' untranslated region. Positive HCV-RNA strand was detected in serum (69%), peripheral blood mononuclear cells (100%) and liver samples (100%). The presence of negative HCV-RNA strand was confirmed using specificity controls assays and was only detected in liver and peripheral blood mononuclear cells samples, (95% and 82%, respectively). No correlation between the presence of negative HCV-RNA strand in peripheral blood mononuclear cells and positive HCV-RNA strand in serum was found, whereas serum HCV-RNA was not detected in patients without negative HCV-RNA strand in the liver. Both positive and negative HCV-RNA strands were found in liver and peripheral blood mononuclear cells of four patients with normal liver histology, and three with minimal changes. Furthermore, the presence of HCV-RNA in serum did not correlate with the alanine aminotransferase values and the histological activity index. These data confirm the existence of replicative intermediates in the liver, not only from patients with histologically proven chronic hepatitis, but also from those with normal liver, suggesting the existence of hepatitis C virus in true healthy carriers.

Liver disease patterns in hemodialysis patients with antibodies to hepatitis C virus

The present study correlated histopathology and diagnostic tests in hemodialysis patients with serologic markers for hepatitis C virus (HCV). Hepatitis C virus infection was found in 65 of 163 patients, as assessed by anti-c100-3 (ELISA 1), anti-c22-3, c33C (ELISA 2), and RIBA 2. Several histopathologic patterns were found in 33 liver samples from HCV-positive individuals: cirrhosis (n = 3), chronic active hepatitis (n = 14), chronic persistent hepatitis (n = 2), isolated hemosiderosis (n = 5), reactive hepatitis (n = 6), and others (n = 3). There was a positive correlation between time from the first aminotransferase peak and histologic damage (P = 0.015). However, the severity of liver disease did not correlate with the intensity of RIBA 2 positivity, mean levels or pattern of aminotransferases elevation, or markers of past hepatitis B virus infection. Moreover, aminotransferases were persistently normal in three patients with severe liver disease and were elevated in 10 patients with only mild changes. In 19 biopsied patients, the presence of plasma HCV RNA was examined by the polymerase chain reaction (PCR), which was positive in 15 of the 19 biopsy specimens. The ability of PCR positivity to predict the histologic severity of the disease was insufficient: four patients with minor liver damage had positive PCR and two patients with significant liver damage had negative PCR. No further correlations of PCR positivity were found with the other biochemical or immunologic markers of HCV infection.(ABSTRACT TRUNCATED AT 250 WORDS)

Treatment with recombinant alpha-interferon of chronic hepatitis C in anti-HIV-positive patients

A pilot study of chronic hepatitis C treatment was conducted in 14 patients (13 had chronic active hepatitis and 1 had liver cirrhosis). All patients were asymptomatic for the human immunodeficiency virus (HIV) type 1 (mean CD4 count of 584 +/- 283 cells/mm3). Patients received 9 MU rIFN-alpha 2A per day for three months. After this, patients received 9 MU three times weekly for three months, 6 MU for another three months on the same protocol, and finally 3 MU again three times weekly for the last three months. After the first month of ALT treatment in 9 patients (64%) returned to normal; a significant decrease in ALT levels was observed with respect to the pretreatment values (mean of 42 IU/l, range 15-75 vs 152 IU/l, range 69-355; P < 0.01). Of the 9 patients who completed the treatment period, 5 had a complete response, and 4 of these 5 continued with normal ALT values during follow-up (sustained response) while the other patient relapsed within one month after cessation of therapy. The remaining 4 patients were non-responders (including one case with a break-through of the response). HCV-RNA was not detectable in 3 of the 5 responders at the end of therapy while during follow-up viral RNA became undetectable in the other 2 patients. 2/4 non-responder patients had detectable HCV-RNA during follow-up. Liver histology improved in all the patients. No changes were observed in the immunological status or HIV infection.

Other approaches to the treatment of chronic viral hepatitis

Interferon (IFN) therapy of chronic hepatitis C has proved useful in recent years. A major goal in the treatment of hepatitis C is the prevention of the high relapse rate after cessation of therapy and the improvement of the response rate in non-responders. Several approaches can be tried. The administration of higher doses of IFN results in an increase in the response rate, but does not prevent ultimate relapse. Similarly, prolongation of the treatment period gives better results, but one third of cases still relapse. In these cases, re-treatment results in ALT normalization once again, but does not prevent subsequent relapse; in non-responders to IFN, re-treatment will not lead to improvement of the response rate. Analysis of the viral genome reveals its presence in serum in spite of the normalization of ALT values. The role of HCV genome heterogeneity in the response to IFN needs to be assessed and it is necessary to find other antiviral agents that are able to clear HCV from all relevant hosts (serum, liver tissue and mononuclear cells).

Immunoglobulin M antibody to hepatitis C virus during interferon therapy for chronic hepatitis C

Testing for immunoglobulin (Ig) M antibody to hepatitis C virus (anti-HCV) as a predictive factor of therapeutic response to recombinant interferon alfa (rIFN-alpha) in chronic hepatitis C was evaluated in 122 patients with IgG anti-HCV. IgM anti-HCV was present in the pretreatment sample of 88% of patients who responded to treatment, including 20 of 21 (95%) long-term responders and 24 of 29 (83%) responders who had relapses after cessation of therapy. In contrast, IgM anti-HCV was present in only 23 of 39 (59%) nonresponders and 22 of 33 (66%) untreated controls (P less than 0.05). The number of cases with detectable IgM anti-HCV tended to decrease in responder patients, which was more evident for complete responders (42%) than for responders who relapsed (72%). During follow-up, the antibody became undetectable in the majority of long-term responders (28% were still IgM anti-HCV positive) but remained detectable in 69% of responders who relapsed (P less than 0.05). No special changes were noted in nonresponder or control patients. Thus, testing for IgM anti-HCV may help to identify a subset of patients who will benefit from rIFN-alpha therapy in chronic hepatitis C.