NEW YORK – Following the start of the European soccer championship in Germany last week, the country is preparing to kick off another large endeavor next month: its national pilot project for genomics-guided precision medicine.
Officially named the "Model Project for Comprehensive Diagnostics and Therapy Selection by Means of Genome Sequencing for Rare and Oncological Diseases," the project is supported with €700 million ($749 million) from the German public health insurance system over five years. Long in preparation, it plans to test the feasibility and utility of decentralized genomic testing in medical care across the country, while collecting data for research and to improve patient care.
Participation in the project has been restricted to university hospitals and their sequencing facilities, leading commercial diagnostic providers, which play an important role in clinical genetic testing in Germany, to fear they will lose out. In the meantime, hospitals have been building up their sequencing and data infrastructure for the project's start in July.
Despite the misgivings, the overall goal of the project appears to have strong support from all providers since it offers, for the first time, nationwide reimbursement from German public health insurers for clinical genomic testing, including whole-genome sequencing, albeit on a trial basis. "We need to bring this established, already-in-use technology into healthcare because it does improve the diagnostic yield for our patients," said Evelin Schröck, president of the German Society of Human Genetics (GfH) and director of the Institute for Clinical Genetics at University Hospital Dresden.
"I do support the model project; it is important," said Elke Holinski-Feder, president of the Professional Association of German Human Geneticists (BVDH) and executive manager of Medical Genetics Center (MGZ), a large private human genetics practice and laboratory in Munich. "For patients who remain unsolved after routine diagnostics, we need additional testing, including genome sequencing, to provide them with a diagnosis," she said. However, this should not happen at the expense of outpatient providers, she added.
After several delays, the model project officially started April 1, according to the final version of the law that governs the project, Article 64e of the legal code that covers Germany's public health insurance system, Sozialgesetzbuch V.
But negotiations about reimbursement rates between the National Association of Statutory Health Insurance Funds (GKV-Spitzenverband) and university hospitals dragged into April, and it took several more weeks for contracts between individual hospitals — the service providers for the project — and health insurers to be finalized.
In addition, the data infrastructure for collecting sequencing as well as clinical data, overseen by the Federal Institute for Drugs and Medical Devices (BfArM), was not yet in place, and service providers had to sign additional contracts with BfArM. Discussions about the type and detail of clinical information to be reported by the service providers also stretched beyond the April start date.
Taking a decentralized approach
The goal now is to start testing patients in July. "We are currently ready — everyone is waiting," said Malte Spielmann, director of the Institute for Human Genetics at University Hospital Schleswig-Holstein, a project service provider for rare disease patients. "We have patients lined up who are excited to be included."
As of now, 18 hospitals and their sequencing labs have qualified as service providers for the project on the rare disease side and 14 on the cancer side, based on stringent participation criteria that included having a human genetics department that is integrated in a university. Final numbers may increase, though, because several applications from additional hospitals are still pending.
"It's really a more decentralized approach, as opposed to the United Kingdom, for example," said Albrecht Stenzinger, deputy director of the Institute of Pathology at University Hospital Heidelberg (UKHD), one of the service providers on the cancer testing side. As a result, he said, "patients have the opportunity to be treated locally near their hometown."
The funding for the project — €140 million per year over five years — will be split equally between rare disease and cancer analyses. "The oncologists had hoped for more, but this was a political decision," said Spielmann, who participated in the negotiations between university hospitals and health insurers.
Initial reimbursement rates will be €8,000 for trio genome sequencing, €6,000 for exome sequencing, and €4,000 for panel sequencing. These rates, which will be reevaluated after two years, cover the entire diagnostic process, including sample prep, sequencing, data analysis and storage, case conferences or molecular tumor boards, and administrative fees. RNA sequencing will not officially be part of the project, but some labs will perform it as an add-on service without reimbursement.
Participating centers will only conduct whole-genome sequencing for rare disease patients. But oncology testing will initially encompass a mix of large gene panel, exome, and genome sequencing, as some molecular pathology labs don't have much experience with exome or genome sequencing yet. Up to 50 percent of cancer samples can initially be analyzed by panel sequencing, which will be phased out after two years in favor of more comprehensive testing, and at least 20 percent of oncology testing must be whole-genome sequencing from the start.
To participate in the model project, rare disease patients need to be undiagnosed and not have undergone exome testing in the past two years, while cancer patients must have failed standard treatments. Eligible patients are discussed at case conferences before being included.
The case conferences, which include several specialists, will act as gatekeepers to ensure that only patients who may potentially benefit from genomic testing actually receive it. "If one of the specialists says: 'I know already what the patient has,' then no genome will be approved," Spielmann said.
The project will also include hospital inpatients, for whom genome testing was not previously covered by health insurance, according to Spielmann. This will enable infants in the pediatric intensive care unit, for example, to obtain rapid whole-genome sequencing, which Hannover Medical School and other centers in Germany have been exploring in a research project called Project Baby Lion.
Each participating university hospital will be allocated a certain number of cases per year for testing, based on the overall number of patients it sees annually. Spielmann said he expects that number to be on the order of 400 to 500 cases, though it might be higher on the rare disease side at the beginning, as there is a backlog of patients who have been waiting for the model project to start. Numbers can also be adjusted over time, depending on how a center performs. The overall goal is to test thousands of patients in the first year, he said.
An important aspect of the project, both for handling reimbursement and for research, is the data platform and IT infrastructure, overseen by BfArM. Sequencing data will be collected by four genome data centers (GRZ), while patients' clinical information — including symptoms, diagnoses, treatment recommendations, and survival information — will be gathered and anonymized by five clinical data nodes (KDK). BfArM has not yet announced the names and locations of the GRZs and KDKs. A spokesperson for the agency said a draft regulation will be discussed in coming weeks by the Bundesrat, one of Germany's federal legislative chambers. "The data platform for the genome sequencing pilot project cannot be put into operation until the ordinance has been issued, but the technical structures for this have already been prepared and processes established," the spokesperson said in an email.
Hospitals will only get reimbursed for testing after depositing a patient's sequencing data to a GRZ and the clinical information to a KDK. "BfArM will get informed by these two locations that we delivered the clinical and the genome data, and once they get the OK, they will send a waiver to the university hospital and the insurance companies, and then we get our money," Spielmann said.
In the short term, the health insurance funds will use the data to evaluate the results of the model project, including diagnostic rates, unnecessary procedures prevented, treatment recommendations, and survival benefits. Based on that, after five years, they will decide whether genomic testing should be reimbursed permanently. "They want to learn to what extent, when, and for which indications broad genomic profiling really has a benefit, and where it might not have a benefit," Stenzinger said. "That's why the whole thing is called 'model project.'"
Longer term, the data will be made available for research, though it might take a couple of years before researchers can start accessing the data. "That's the larger benefit beyond just rendering a molecular diagnosis and facilitating personalized treatments," he said. Similar to a clinical trial, patients can opt out of research use of their anonymized data if they wish.
Private providers excluded
Not everyone is happy with the current setup of the project. Outpatient diagnostic providers cannot become service providers under the eligibility criteria, which require integration with a university that conducts research and training, even those with significant experience and capacity for clinical exome and whole-genome testing.
Dirk Biskup is co-CEO of CeGaT, a large genomic testing laboratory based in Tübingen. "Of course, CeGaT would like to participate in the project, like many nonuniversity labs do," he said in an email. "CeGaT runs one of the largest next-gen sequencing labs in Europe. It is fully accredited and processes more patient samples than any university lab in Germany," he said. "There is absolutely no good reason to exclude a highly professional lab such as CeGaT."
According to Biskup, the vast majority of genetic diagnostics in Germany is currently provided by outpatient medical genetics practices and their laboratories. "Many of these specialized practices are larger, more experienced, and more professional than university labs," he said.
Founding members of the National Alliance for Rare Genetic Diseases (NASGE), for example, an association of medical geneticists in Germany that offer outpatient diagnostic services, provided almost 37,000 clinical exomes or trio exomes and almost 500 genomes in 2023, according to the organization's website.
Given that the model project is funded by taxpayer money, excluding certain providers is problematic, "if not illegal," Biskup added. "From my point of view, the only criteria for eligibility should be quality (accreditation) and the ability to deliver the defined set of NGS data, medical reports, and patient care. Every provider able to deliver the defined medical services at an accredited quality should be eligible."
Holinski-Feder from BVDH and MGZ echoed those sentiments. According to her, not all universities necessarily provide high-quality genomic testing. "Some do, but by far not all. This is not fair, this institutionalization of quality, because other groups are being excluded," she said.
Also, the negotiated reimbursement rates for the model project are considerably higher than what private diagnostic labs currently obtain for the same kind of testing. Holinski-Feder said her practice currently receives less than €2,000 in reimbursement for an exome test and the same amount for a genome test, which the center occasionally runs for unsolved patient cases. "This means that if you do genetic diagnostics as an outpatient provider, you get paid much worse than being part of the model project for essentially the same service," she said.
In addition, public health insurers in each federal state decide how much of their total available funding — the so-called "morbidity-related total compensation" — is dedicated to human genetics services each year. When the allotted amount is used up, reimbursement rates are cut. "The size of the pie remains the same; the spoons get smaller," she explained.
If the model project provided additional funding for genome sequencing services, "we could live with all of this because universities need more money, et cetera," she said. "But the funding for the model project is supposed to come from the same pot of money that pays for outpatient services," she added, potentially resulting in severe cuts for providers like her own center.
As evidence, she pointed to the law that governs the model project, which states that "the morbidity-related total compensation is to be adjusted for the services provided within the framework of the model project."
If health insurers indeed start reducing the amount of funding available for human genetics outpatient services, because these will now in part be provided by the model project, MGZ and other human genetics practices plan to file a lawsuit, Holinski-Feder said. "If health insurance wants to provide additional financial support, they should go ahead and do so, but it cannot come from our pot of funding. They cannot say: 'We're doing a great project, but someone else should pay for it.'"
Members of participating hospitals argue that the clinical utility of whole-genome sequencing remains to be demonstrated and that university hospitals with their many subspecialties are the best equipped to do so. "Once they show that this is indeed beneficial, there will likely be a 'normal' way to reimburse this, and then everyone can participate," Schleswig-Holstein's Spielmann said.
"It does make sense to start with academic institutions because they have the full infrastructure in terms of precision medicine; it's not only about the diagnostics per se but also about the right treatment, follow-up of patients, and so on," UKHD's Stenzinger said. "Going forward, you will certainly need those private institutions, as well."
Public health insurers are also afraid that private labs might provide genome sequencing tests unnecessarily, Spielmann said, since they have an economic incentive to do so. "Some of the private institutes are extremely capable and extremely knowledgeable; they know what they're doing, but they are mainly financially driven," he said. "In a system where money and resources are limited, [health insurance is] a bit scared of this because it has happened before that the private sector sequenced too many [patients]," exhausting available funds.
However, restrictions on providers will not prevent genome sequencing, which has been shown in many studies to improve diagnostic yield for rare disease patients, from becoming the test of choice eventually, according to Holinski-Feder. "You can't stop the world from turning," she said. "Eventually, we will all do genomes, this way or that way."
A boon to Illumina
No matter whether university hospitals or private practices will do the testing, one company stands to gain: Illumina. "For us, the model project is very exciting," said Sven Schaffer, global head of medical affairs at Illumina, who is based in Hamburg. "I think we can learn a lot from this, not only in Germany: how feasible it is to bring whole-genome sequencing into a hospital and close to the patient."
All participating university hospitals run Illumina sequencing technology already, he said, and would need NovaSeq X or NovaSeq 6000 high-throughput sequencers in order to run whole-genome sequencing tests. "Some of them are still in the process of purchasing high-throughput instruments, and others made the decision a while ago," he said. In addition, there is a lot of investment in ancillary equipment, as well as infrastructure for data processing and data storage.
While most participating human genetics centers were already equipped with NovaSeq instruments, which have been mainly used in research projects so far, he said, some of the pathology labs currently still run large panel tests on NextSeq sequencers. "They could potentially postpone the investment in a high-throughput sequencer for whole-genome sequencing," he said, but will eventually need to upgrade.
According to Schaffer, "the readiness for whole-genome sequencing in oncology is a challenge for all laboratories, not only in Germany." One reason it is not well established in a routine clinical setting for cancer is because it requires fresh frozen samples, he added, which can be difficult to obtain.
Schaffer said he expects the project to take off slowly. "This will take some time, but it will start. The limiting factor currently is the lack of data infrastructure," which is taking more time than expected to build. Other challenges include limited standardization and guidelines for clinical whole-genome sequencing and the need for labs to validate the equipment. "Finding the resources to do all this in a clinical setting is probably a burden for the hospitals," which are strapped for cash, he said.
The model project does not prescribe a specific sequencing technology, though, and several university hospitals, including in Schleswig-Holstein and Tübingen, are already experimenting with long-read technologies from Pacific Biosciences and Oxford Nanopore Technologies as a complement to short-read sequencing and with the intent to use them in the model project. "For rare diseases, this will happen starting in the second year or so," Spielmann predicted.
University sequencing labs gearing up
Meanwhile, sequencing labs at participating university hospitals are preparing themselves for the July start date. Nicole Pfarr, head of molecular pathology diagnostics at Technical University of Munich (TUM), said her lab had to establish and validate clinical whole-exome sequencing for cancer testing. As part of this, they compared exome kits from Integrated DNA Technologies (IDT) and Twist Bioscience and found them to deliver comparable results.
Pfarr's lab also participated in an exome sequencing pilot study organized by the German Network for Personalized Medicine (DNPM), in which 21 centers participated, including all pathology sequencing labs that are part of the model project. For the pilot, participants sequenced FFPE samples, testing exome kits from IDT, Twist, Agilent Technologies, and Qiagen. "We do not expect to get fresh or fresh frozen tissue from all the patients, but we will have a lot of patients where we have only FFPE material," Pfarr explained. Somewhat to her surprise, her lab's FFPE exome results were very similar to those from fresh frozen material. "It worked quite well, even when we had a couple of very compromised samples," she said.
At present, pathology labs need to report on 490 genes from the exome data, Pfarr said, as well as conduct other analyses, such as tumor mutational burden testing. They also need to do germline testing for about 40 genes, often in collaboration with their human genetics departments. There are ongoing efforts to harmonize reporting for the model project, so all patients receive the same type of report, she said.
According to Pfarr, a big challenge will be variant interpretation, including variants of unknown significance, and using the results to guide therapy selection. "We do not know what the oncologists will make of all this information," she said. In the beginning, not all patients will likely see a benefit from their testing results, "but maybe in the future, if new therapies are coming, if new studies are on the horizon, patients might be included in those studies," she said. In addition, follow-up clinical data on patients will provide new insights into the mechanism of their disease.
Further, reimbursement could be a complex process, Pfarr said. Patients eligible for the project first receive a number from BfArM to anonymize their identity, she explained. After that, the institute has 28 days to conduct the genomic testing, bioinformatic analysis, and molecular tumor board reporting. Subsequently, it needs to upload the sequencing data to a genomic data center and the patient data to a clinical data node, and only then can it be reimbursed by health insurance. The entire process may take weeks to months, Pfarr predicted, so even if her institute starts analyzing patients in July, it may not be reimbursed until October or so.
Also, reimbursement needs to be split between different entities who all want a piece of the pie, including the clinic and the pathology institute. For a panel test, her institute may end up only receiving €2,000, or 50 percent of the total payment. "We will just get enough money to cover the material costs, and maybe the technicians or scientists," she said.
In the first year, she expects her institute, which collaborates with Klinikum rechts der Isar, TUM's university hospital, to test at least 350 patients, though she said that number will likely grow to about 700 to 800 samples per year later in the project.
While the start of the model project is close, it may take a while before it gathers momentum. "We're probably going to see a rough start and experience a learning curve over the next couple of years," Stenzinger predicted. "I doubt that the data infrastructure will be up and running from day one. These things will certainly take a while and will probably be developed while the whole thing is already in fully swing, if you will."
"I will be very relieved when it's up and running," Spielmann said. "It's been stealing all of our time, but it's important, and I think we are in a very good position to do something good."