The 100,000 Genomes Project was a landmark initiative in the United Kingdom, evaluating the feasibility and benefits of WGS in cancers and rare diseases. This report summarises the experience of WGS for cancer patients who participated in the Project within the West Midlands GMC, the GMC that had recruited the most participants for the Project and covering one of the most diverse populations in the country (Fig. 1). We demonstrated rapid improvement in recruitment and the quality of the samples received during the Project (Fig. 2; Supplementary Fig. 3). Unsurprisingly, the exponential increase in recruitment was accompanied by an increase of other sample issues, including insufficient samples. Of the 4851 sample sets included in the study, 3173 were successfully sequenced, representing 3067 participants of disparate cancer types.

Approximately 60% of participants (1811/3067; 59.0%) were clinically excluded as GTAB discussion would not have changed clinical management (including those who had died, with early-stage cancers or poor fitness). This rate of pre-GTAB exclusion varied significantly between cancer types (Table 1 and Supplementary Table 8), especially in cancers that present with diagnostic challenges or poor fitness at presentation, such as cancer of unknown primary and lung cancer [20, 21]. Conversely, the high rate of exclusion of breast and colorectal cancer participants likely represents a high proportion of curative intent treatment at presentation.

The 100,000 genome project adopted an unselected approach to sequencing, and did not exclude patients based on stage, grade or prognosis. Universal sequencing for cancer patients provides opportunities, including identification of otherwise unrecognised germline variants and contributing to the understanding of the biology of malignancy [6]. However, as demonstrated by our data, a potential weakness of this strategy is that a significant proportion of patients will have sequencing that will not change their cancer management. As a result, the approach presented here may under-estimates the clinical utility of WGS when compared with an approach applied strictly to patients with a permissible performance status and advanced disease.

Six in ten participants discussed at GTAB had recommended actions (59.1%, 548/928); four in ten participants discussed had potential treatments or clinical trials recommended to them (40.4%, 377/933; Table 2), including in cancers with limited therapeutic options such as glioma (50%, 34/68) and hepatobiliary cancers (49%, 50/102). Unfortunately, only a small proportion of potential treatments or clinical trials were followed at the time of last follow-up (5.1%, 25/491). Whilst this is clearly disappointing, it is consistent with reported rates of involvement of cancer patients in clinical trials [22], which is further reduced by socioeconomic deprivation and in ethnic minority groups [23]. An additional challenge for our population was likely presented by the COVID-19 pandemic [24], which may have restricted access to clinical trials and novel therapies. It is possible that with longer follow-up more of our participants may access stratified medicine options based on the WGS data, particularly at times of disease relapse or recurrence.

The Project is the largest population-based genomics medicine trial with an overarching aim to introduce WGS within a publicly funded health system. Initial studies using the data from the Project have focused on the genomics and clinical features of specific cancer types [25,26,27,28]. Others have explored the potential clinical utility in rare cancer types, including sarcoma and cancers in children [4, 25]. The recently published pan-cancer evaluation of successfully sequenced specimens [6] has used data linkage approach to explore the relationships between tumour genotypes with associated survival. However, it lacked data associated with the clinical decision-making process, nor possible outcomes from patients’ perspectives to support informed shared decision making (Fig. 1). The small but important group of patients with cancers of unknown primaries and haematological cancers were also excluded.

Prendergast and colleagues reported similar rates of WGS being performed from sarcomas (62%) [25]. Prior chemotherapies [25, 29], sample types and disease-specific challenges (e.g., low-grade disease with few cells, or high-grade disease with necrosis) were previously associated with higher failure rates of genetic testing. In addition, treatment-pathway-related factors, such as timing of referral (post-diagnosis) and out-of-hours surgery, have been reported to influence success rates [25].

Similar to a recent report on childhood cancers [4], in which 1 in 5 (22.2%; 8/36) participants undergoing WGS were recommended novel therapeutic opportunities, 3 in 10 (29.8%; 252/846) of our participants whose results were discussed at GTABs were recommended potential treatments or clinical trials (Table 2).

This study evaluated the clinical processes supporting the implementation of WGS during the Project. It has highlighted the barriers for all cancer types, variations in success rates of WGS and the associated availability of therapeutic opportunities between cancer types. This report complemented the recently published pan-cancer mutational characteristics of the Project [6] with GTAB outcomes of individual participants to demonstrate how WGS has influenced clinical decision making.

Although we identified improvements of quality control during the Project, it is unclear whether any pathway changes were implemented (e.g., removal of formalin from theatre), as previously described [25, 30]. The lack of detailed demographic data, histological subtypes and details of prior treatments precluded further evaluation of risk factors associated with WGS failure and potential strategies to maximise its potential clinical utility.

Our data have not formally evaluated the frequency of diagnosis modifications based on WGS, although and no recorded outcomes from GTAB suggested that diagnosis had been altered. Respectively, 3.0% and 16.7% participants with sarcomas [25] and childhood cancers [4] had their diagnoses refined or changed by their WGS results.

A significant disadvantage of the study was the turnaround time for WGS, caused by the evolution of infrastructure needed to deliver it. However, time to sequencing fell from 16 weeks at the beginning of the Project to 4 weeks at the end, suggesting that building infrastructure and experience facilitates more efficient delivery of WGS. Another weakness of our analysis is that clinical data such as cancer stage was not collated, we are therefore unable to provide a data on the relative utility of WGS in patients with advanced disease.

It is important to consider whether WGS is the optimal strategy for universal cancer sequencing. WGS provides comprehensive genomic sequencing, representing a more diagnostically rich approach [31, 32]. In addition, the importance of rich WGS data for cancer research must not be understated [6]. WGS offers the potential to institute standardised pan-cancer sample processing and bioinformatics pipelines, which is attractive for a nationalised healthcare system. A potential disadvantage of WGS is that this data risks overwhelming clinical teams with unexpected results. In the 100,000 Genomes project, this was offset by use of virtual panels, which were then interpreted within the clinical context by GTAB teams. Unlike traditional panel-based sequencing, virtual panels may evolve with scientific understanding [33], allowing for re-analysis of samples, and for evolution of testing to meet service requirements.

Disadvantages of WGS must be acknowledged, with one drawback being time to results and cost-effectiveness compared to alternate technologies [34,35,36]. Whole-exome or panel sequencing could offer a route to quicker turnaround times, with more rapid turnaround of results potentially allow more patients to access therapies before clinical deterioration. However, as shown here, time to sequencing does fall with establishment of infrastructure for WGS within a healthcare system. A further important consideration is that WGS necessitates either sequencing at either lower depth per gene or committing to prolonged sequencing time [36]. Finally, WGS is poorly validated in FFPE. Successful sequencing is therefore dependent on fresh or fresh-frozen sample, with extraction of relatively large amounts of high-quality DNA [9]. A successful WGS programme therefore requires establishment of robust sampling protocols, with cross-team collaboration [9]. However, the success of the 100,000 Genomes Project demonstrates the feasibility of establishing such a programme at a national level. Clearly the ideal sequencing modality for cancer patients will be influenced by the healthcare system, resources and local expertise.

This study provided an estimate of overall rates of successful WGS that led to clinically actionable recommendations. This will improve counselling of patients in a population-based setting with a mixture of different histopathology pathways from both secondary and tertiary hospitals. Notably, recommendations for potential therapeutic options were made in 49.0% (50/102) of participants with hepato-pancreatico-biliary cancers discussed at GTAB and 50% (34/68) of adult participants with glioma discussed at GTAB. This suggests that WGS may provide therapeutic avenues for patients with currently very limited therapeutic options. Targeting these cancer types to establish robust molecular profiling pathways will help achieve potential therapeutic benefits of WGS. In contrast, the management of CUP remains an unmet need. Only one participant with CUP had a potential therapeutic option suggested by GTAB. Of 17 participants recruited, 10 (58.8%) died or were clinically excluded before results interpretation and 5 (29.4%) had no actionable variant. A standardised pathway is needed to expediate diagnosis and management in this challenging group.

It is acknowledged that the genes within Domain 1 are broad, containing genes that are not classically viewed as clinically actionable with current standard of care therapy options. However, these domains have been selected based on expert consensus [10]. Identifying patients with genetic variants that allows matching of patients to stratified medicine trials, providing potential additional treatment lines and providing an opportunity to increase recruitment to these valuable trials.

Previous work highlighted the optimistic view of participants and health professionals on the potential of WGS to improve therapeutic options [37,38,39]. However, the lack of matched agents with targets remains a hurdle. Careful comparisons of key molecular pathways will generate new hypotheses to evaluate the benefits of targeted therapy in different cancer types. Expanding access to stratified medicine through clinical trials is clearly a priority. For cancer types that have high clinically actionable variant rates and few therapeutic/clinical trial options, tumour agnostic trials are urgently needed to improve outcomes. Whilst large multi-agent, multi-marker studies have been performed, in many cancers the ability to match agents to specific genotypes remains limited. The Cancer Research UK funded DETERMINE study will undertake genotype matched therapy in rare tumours where the therapy is not licensed for that indication to maximise potential benefits of WGS (Clinical Trial registration ID: NCT05722886).

Further research is needed to systematically test methods for improving WGS pathways and implementation of GTAB recommendations. Facilitating use of diagnostic pre-treatment specimens that yield good quality nucleic acids is crucial for implementation of genomics [25, 29]. Moreover, evaluation of patient characteristics (including socioeconomical factors, frailty and ethnicity) will identify good practice and barriers to implementation of genomics in different cancers.