Accurate detection of tumor fusion gene

Gene fusion caused by chromosome rearrangement is a common phenomenon in tumors and plays a very important role in the occurrence of tumors, accounting for about 20% of the incidence of human cancer. At present, there are many kinds of tumor targeted therapy drugs that inhibit the fusion gene at home and abroad, including imatinib /BCR-ABL 1, crizotinib /EML4-ALK, and lalotinib /NTRK fusion.

However, the frequency of fusion genes in different cancers varies greatly, and many fusion genes only exist in specific cancer subtypes. For example, the positive rate of ALK fusion gene in non-small cell lung cancer (NSCLC) is 3% ~ 7%; ROS 1 fusion gene only accounts for 2% of lung cancer patients. The average positive rate of NTRK gene fusion in all solid tumors is about 65438 0%, but it is as high as 90% in some specific tumors.

Opportunities and Challenges How to Realize the Accurate Detection of Fusion Genes?

Therefore, accurate detection of fusion genes can not only help clinicians to diagnose and classify cancer, but also provide necessary information for follow-up treatment. At present, fusion gene diagnosis mainly includes fluorescence in situ hybridization (FISH), IHC and so on. However, these detection methods usually have low resolution and flux, and also rely on the judgment experience of inspectors, and are often only suitable for known fusion subtypes, so their application is limited, which often leads to a long diagnosis process and high cost. As shown in fig. 2, the probe can still bind to the target region because the missing region is small. Therefore, the fusion of GOPC-ROS 1 was not detected in the FISH results, and there was a false negative. It has been reported that the sensitivity of IHC in detecting NTRK3 fusion is decreased, and the sensitivity of NTRK 1 and NTRK2 is 97% and 99% respectively, while NTRK3 is only 79%. In addition, its specificity is also related to the tumor type. [5]

Although traditional detection methods will still be an indispensable means of tumor fusion gene diagnosis, at the same time, a variety of new technology platforms are gradually entering the clinic (table 1). For example, the fusion gene detection method based on hybridization/fluorescence or mass spectrometry technology avoids the subjective judgment of technicians to a certain extent, is more digital, and has economic advantages while improving throughput; On the other hand, more than 10000 tumor fusion genes have been found in clinical research at this stage. Although all the above techniques have high detection sensitivity, no new fusion subtype can be found, and there are still some limitations. In addition, although NGS-based anchored amplicon technology is convenient and fast, it does not need prior knowledge of fusion, but it has a slight disadvantage that it is easy to produce false positives, and it is necessary to filter out fusion genes with low confidence by using biological information pipeline.

Multi-dimensional comparison of èRNA-Cap has great potential for improving the detection of clinical fusion genes.

Notably, in recent years, DNA-Cap technology based on hybridization has been increasingly adopted by clinical laboratories because of its advantages of economy, independence from prior knowledge of fusion, ability to find new fusion subtypes, versatility and high sensitivity to genome and free nucleic acids. ?

At present, many DNA-Cap adjoint diagnosis products based on NGS technology have the ability to detect fusion genes, such as MSK-IMPACT. The whole process is as follows: designing a probe covering the intron region of a functional gene, hybridizing the probe with a pre-library, capturing and sequencing, analyzing and counting chimeric readings, and finding fusion events. However, using DNA-Cap alone to detect fusion genes also has some limitations:

1.？ sensitive

The sensitivity of NGS technology is directly related to the coverage, and some introns are very long, such as the intron (~ 193kb) in NTRK3. If it is covered, it will further increase the cost. In addition, some intron regions can't be designed with probes even if necessary, because they contain a large number of repetitive sequences, which will reduce the target and make it difficult to compare. For example, intron 3 1 of ROS 1, as the main fusion region (CD74/SLC34A2/SDC), contains more than 90% repetitive sequences. Therefore, if the fusion breakpoint involves uncovered intron regions and the capture probe cannot cover these regions, the detection sensitivity will be affected. For example, in some FDA-approved tumor groups, hotspot intron partial coverage and companion coverage (NTRK3-ETV6) are usually used for NTRK fusion, which may lead to the risk of missed detection (Figure 3).

2.？ characteristic

Different analysis software leads to different false positive rates, so some studies use a variety of fusion analysis software to improve the specificity [7]; What is more noteworthy is that it is difficult to confirm whether the fusion gene is expressed or not by chromosome rearrangement or gene fusion at the genome level. In a DNA/RNA study based on NGS, MSK-IMPACT Panel found 23 NTRK fusion positive samples, of which 265,438+0 were confirmed to have fusion transcript expression, and 2 negative samples had neither fusion transcript expression (Archer RNA detection) nor fusion protein (IHC), which shows the importance of verifying genome fusion [8].

Although some studies have used genome-wide sequencing to detect fusion genes, after pathological examination, most clinical samples of tumors are FFPE, and their RNA is generally partially degraded, with low quality. In the NGS-based RNA-seq technology, mRNA-seq has great limitations. In the study of breast cancer fusion genes using mRNA-seq, only 45.7%(32/70) of the fusion genes can be verified by RT-PCR and Sanger method [9]; However, when ribosome removal method is used, because of the existence of mRNA precursors, the intron content increases, the exon ratio is relatively low, and the effective data is low and unstable (1%-30% exon ratio), because of RNA degradation, rRNA removal is insufficient. In addition, non-target genes in tissues will further reduce the data proportion of target exons. In other words, genome-wide sequencing can be understood as a natural WES, but due to the differential expression of genes, the transcript coverage will be uneven, and it will be limited by experimental techniques (ribosome removal, reverse transcription, etc.). ) and biology (mRNA precursor), and ultimately a large amount of data may be needed to improve the sensitivity and stability of RNA-seq in detecting fusion genes (Figure 4).

Comparatively speaking, RNA-Cap technology can overcome the above bottlenecks in the detection of fusion genes. Firstly, the probe designed for the transcript avoids the interference of intron region, effectively improves the sensitivity, and makes the panel smaller and more economical. Secondly, rRNA pollution and interference of introns and non-target genes are effectively avoided, thus improving the validity of data, the depth of target region and the detection sensitivity (Figure 5); More importantly, RNA-Cap can evaluate the function (in-frame) and expression of the fusion gene while discovering the fusion gene, and verify the fusion gene for the second time. Because of this, MSK-IMPACT now uses DNA-Cap and RNA-anchored amplicon technology to detect selected patients. ?

At the same time, use DNA&; amp； amp； amp； amp； amp； amp； amp； amp； amp； amp； amp； The genomic variation and mutant gene expression were evaluated. The two-dimensional detection method of RNA*** capturing fusion genes can be said to be a microcosm of the development of precision medicine. From "one medicine and one examination" to "small tumor panel/large tumor panel" to "WES", tumor-related diagnostic technology is progressing at an unprecedented speed. It is believed that with the maturity of technology and the reduction of sequencing cost, high-throughput sequencing will better serve patients.