Improving Genomic Profiling and Reducing Time to Cancer Treatment Via Targeted Use of Endoscopic Ultrasound

Overall aim:

To investigate whether the novel FNB needle provides a better diagnostic accuracy than a standard FNA needle in EUS-B guided cancer diagnosis.

Design:

Prospective multicenter randomized control superiority study.

Primary study objective:

Comparing the ability to provide comprehensive phenotyping from microscopy histology/cytology to the molecular level to allow for optimal and early personalized cancer-specific therapy of FNA and FNB from tumors and lymph nodes adjacent to the esophagus/stomach using EUS-B in patients referred with suspected cancer.

The main focus are common deep-seated malignancies: NSCLC and lymphoma.

Study description:

Consecutive patients referred due to suspected cancer with enlarged (at least 10 mm in the shortest axis) and/or FDG-avid lymph nodes or other lesions adjacent to the esophagus/stomach (e.g. suspected liver metastasis), thus with an indication for EUS-B, will be randomly assigned for tissue sampling either with a standard 22G FNA needle, or the novel 22G crown-cut FNB needle.

280 patients with suspected lung cancer will be included and inclusion will end when the targeted number of 254 patients (127 in each group) with a final diagnosis of NSCLC is reached (see under statistics).

Likewise, 82 patients with suspected lymphoma will be included until the targeted number of 74 patients (37 in each group) with a final diagnosis of lymphoma is reached (see under statistics).

Needles:

The standard needle for EUS-B is a 21G or 22G FNA-needle (fine needle aspiration), but new needles have been developed to obtain more histological material for immuno-histochemistry and mutation analyses. Larger biopsy specimens increase the diagnostic yield and allows for more mutation analyses, yet sample quality is not only determined by volume: larger tissue volume is obtained with a 19G versus a 22G FNA needle but with identical diagnostic yield.

Tissue volume seems also increased with EBUS using a 22G needle with a novel three-point crown-cut design (Franseen needle design) compared to a standard 22G FNA. However, the diagnostic yield did not increase, and one of the reasons addressed was the inability to penetrate the tracheal/bronchial cartilage (which is not at problem in EUS-B). Safety and adverse events did not differ from FNA.

Crown-cut needles belongs to the third generation of fine needle biopsy needles (FNB). Within the third generation of FNB needles, studies comparing the different designs found that FNB needles with a multifaceted tip (crown-cut and fork-tip needles) were more superior than other designs of the same generation.

No study has investigated the use of the FNB needle in EUS-B. On this background, in the EUS-B study the investigator describe a systematic evaluation of the diagnostic superiority of a novel 3rd generation 22G FNB needle (SonoTip TopGain®, MediGlobe, Germany) over standard FNA, as well as evaluating its safety in EUS-B.

Study conduct:

Recruitment and informed consent:

The investigator will recruit participants to the study among patients referred to the study sites with suspected cancer. Upon planning the diagnostic procedure, the doctor assesses if the patient can be included in the study. Here, the patient will receive oral information about the study by a healthcare professional who is familiar with the study. The conversation will take place in undisturbed conditions, where the patient can bring a co-sitter.

The patient will the receive a written information about the study, as well as a document explaining research participants rights in a health science research project as well as contact information about the research responsible doctor.

24 hour consideration period is given to the patient who will be able to contact the research responsible doctor with questions about the study.

The patient can be included in the study once both the healthcare professional and the patient have signed the written informed consent document.

The patient have will have the ability to withdraw from the study at any time and without any consequences on further diagnostic investigation or treatment.

Recruitment will begin in Region Zealand (Næstved and Roskilde) and once the study is established, a cooperation agreement as well as data processing agreement will be made with Odense University Hospital.

The EUS-B procedure:

The patients will be allocated to one of two groups based on either suspected lung cancer or deep-seated lymphoma. In each group, randomization will be performed in a 1:1 ratio using an online password-protected computer-generated system (RedCap). Moreover, randomization will be stratified based on cancer history, to ensure that there will be as many with suspected first time lung cancer or lymphoma and suspected recurrence/progression of a known cancer in each group.

If the patient is found suitable for inclusion and informed consent is obtained, the patient is immediately randomized to sampling with either FNA or FNB prior to the procedure.

The diagnostic work-up will be performed under conscious sedation as an out-patient procedure. After bronchoscopy (and EBUS if indicated), EUS-B is performed by sliding a linear EBUS echo-endoscope down the esophagus.

Sampling technique will be identical for both needles. Sampling, using an inner-stylet, will be performed by minimum two needle-passes per station using a fanning technique with suction. However, number of needle-passes might change if the operator finds it indicated, and it will be noted. Aspirates will be transferred to a dedicated container containing formalin, and at least one smear is produced on a glass slide. Samples are marked with patient ID and lesion location.

The pathologist will be blinded to group allocation. Only the study coordinator will have access to the randomization module.

Rapid on-site evaluation (ROSE) is not available at the study sites and will therefore not be utilized in the present study.

The diagnostic procedures will be performed by endoscopy experienced pulmonologists (minimum 200 EUS-B performed) from the respiratory department of each of the following Hospitals:

• Zealand University Hospital, Næstved
• Zealand University Hospital, Roskilde
• Odense University Hospital, Odense

Biological Material:

Needle biopsies of suspicious lesions are sent as per normal clinical practice to the Department of Pathology, where the material will be examined for the presence of cancer cells and genomic analyses will be performed when Non-Small Cell Lung Cancer (NSCLC) is found.

Tissue material will as usual practice be received and analyzed at the Department of Pathology, where they have a Diagnostic Biobank to store tissue samples. This is not specific to this research study, and the investigator will not store any tissue after analysis.

Specimen processing:

All material will be prepared as per clinical standard, where tissue micro cores will be processed as cell blocks with formalin fixation and paraffin embedding. Smears are airdried for May-Grünwald-Giemsa staining. Each cell block will be stained with haematoxylin and eosin, and when relevant immunohistochemical stains specific for lung cancer will be used. In case of other cancer, supplementary stains will be used according to standard regimens. Tissue is additionally sent for flow cytometry in dedicated glass tubes when lymphoma is suspected.

Diagnostic accuracy:

Diagnostic accuracy of FNA and FNB will be compared and estimated using the equation "diagnostic sample/all samples". Specimens will be evaluated according to a local modified version of the Papanicolau classification of pulmonary specimens and categorized as either:

1. Non-suitable II. Negative for malignancy III. Atypical cells (e.g. inflammatory or reactive changes) IV. Suspicious for malignancy V. Malignant

Mutational analysis:

Following tumor content evaluation of hematoxylin and eosin-stained slides, and NSCLC is concluded, relevant regions are macrodissected and subjected to a standard genomic DNA extraction procedure using the GeneRead DNA FFPE Kit (Qiagen). DNA-concentration is estimated using a Qubit Fluorometer. All methods are carried out in accordance with the manufacturer's guidelines with at least 20 % neoplastic content (estimated by pathologist) in each tissue sample. Next-Generation Sequencing (NGS) using the available NGS assay at the Department of Pathology will be performed on all NSCLC patients regardless of subtype.

The focus of the investigatior will be on the lung cancer associated biomarkers e.g. ALK, BRAF, EGFR, ERBB2, KRAS and MET.

Regarding lymphoma, analysis of tissue material will follow the normal clinical practice at the study sites, where subtyping is primarily based on immunohistochemistry and flowcytometry. The above mentioned NGS assay is performed in cases where tissue material is sparse. NGS is then necessary for classification of lymphoma subtype.

Statistics:

For the suspected lung cancer subgroup:

Sample size calculation was based on expected difference in prevalence of successful comprehensive targeted genomic profiling of 13% (77% in FNA vs 90% in core needle biopsy, CNB). With 80% power (1-beta), significance level (alpha 0.05) this yields 254 patients. It is expected that 10% (26 patients) will have other diagnoses, so the number of included participants has to be 280 patients to diagnose 127 with NSCLC in each group (FNA vs FNB).

For the suspected lymphoma subgroup:

Sample size calculation was based on expected difference in diagnostic accuracy of 31.30% with transesophageal sampling of deep-seated lymphoma (91.30% in EUS-FNB and 60% in EUS-FNA). With 90% power (1-beta) and significance level (alpha) 0.05 this yields 74 patients.

It is expected that around 10% (8 patients) will have other diagnoses, so the investigator must include 82 patients to diagnose 37 patients in each group.

Descriptive and analytical statistics will be performed using dedicated statistical software (STATA18.0; StataCorp®, Texas, US). Categorical data will be described as number (n) and percentage (%), and continuous variables as median and range. Intergroup differences in categorical variables will be analysed with Chi2-test or Fisher's test (if applicable), and differences in continuous variables with Mann-Whitney U-test.