Unknown Primary Tumors

This study will collect, annotate, and sequence biospecimens (blood, tissue, urine, saliva and surgery drainage) from patients across different cancer types to detect molecular residual disease (MRD). Imaging scans and clinical data will also be gathered. This will allow for early cancer interception, and hopefully prolong relapse-free survival across tumor types. Results of ctDNA testing will be provided for clinical decisions and to determine eligibility for other linked interventional interception therapeutic studies, each of which will have a separate protocol.

This study aims to initially utilize machine learning on pan-cancer DNA methylation data from public databases to construct a DNA methylation classification model (PaCIFiC-CUP, pan-cancer integrated fingerprinting classifier of CUP) for diagnosing various types of cancer, particularly the primary site of cancer of unknown primary. The goal is to achieve diagnosis of cancer pathology type by analyzing the DNA methylation patterns of cancer specimens, thereby guiding subsequent precision treatment for cancer.

This study will investigate the effects of atezolizumab on select cancer types in people whose analysis of tumour DNA and RNA indicates they may be sensitive to atezolizumab. This study aims to determine if the information from the cancer genome analysis corresponds with the effects of atezolizumab on individuals and their cancer. This is a Phase 2 study, which is undertaken after preliminary safety testing on a drug is completed, and will involve approximately 200 participants. Participants are assigned to one of 8 cohorts based on their primary tumour type: breast, lung, gastrointestinal (GI), primary unknown, genitourinary (GU), sarcoma, gynecological, and 'other' cancer types. Participants in all cohorts will receive the same dose of atezolizumab (1200 mg every 3 weeks). In the first stage for each cohort, 8 participants will be enrolled and if no participants respond to treatment, enrollment to that cohort will be closed. If 1 or more participants respond to treatment, up to 16 additional participants will be enrolled to that cohort. Participants continue on treatment until they no longer may benefit from the treatment or they decide to stop treatment.

During the last decades hematologists have excelled at improving and refining the classification, diagnosis, and thus ultimately the therapeutic decision-making process for their patients. This continuous evolution proceeded in parallel to seminal discoveries in basic science such as FISH, PCR and NGS. So far, the current WHO classification serves as reference to diagnostic decision making and is largely based on 5 diagnostic pillars: cytomorphology of peripheral blood and/or bone marrow smears, histology and immunohistochemistry of bone marrow trephine biopsies or lymph nodes, immunophenotyping, chromosome banding analysis supplemented by FISH analysis, molecular genetics including PCR and targeted panel sequencing via NGS. This leads to a swift diagnosis in 90 % of all cases. The leftover 10 % remain a challenge for hematopathologists and clinicians alike and are resolved through interdisciplinary teams in the context of specialized boards. With the advent of high throughput sequencing (mainly WGS and WTS) the possibility of a comprehensive and detailed portrait of the genetic alterations - specifically in challenging cases - has become a realistic alternative to classical methods. In SIRIUS the investigators will prospectively challenge this hypothesis to address the question of how often a better or final diagnosis can be delivered by WGS and/or WTS and if unclear cases can be efficiently resolved.