Summary

Endoscopic ultrasound fine needle aspiration (EUS FNA) is an established and recommended technique for diagnostic of solid pancreatic masses. The accuracy of the technique depends on the operator experience, lesion type and location, type of procedure sedation as well as procedure related technique factors (presence of elastography or contrast enhanced imaging, needle diameter, presence of stylet, use of suction and type of suction, the number and method of "to and fro" movements, the number of passes and the presence of a cytopathologist in the examination room).

The relationship between the "to and fro" movement and the EUS FNA yield in solid pancreatic masses has only been explored in the literature in a subjective fashion, without accurately measuring the needle acceleration.

Recently, a simple electronic sensor device connected by Bluetooth to a phone, has been proposed for teaching and research purposes. Among its sensors, it includes an accelerometer which can measure the instant scalar acceleration of an object and transmit it to the connected phone. By attaching this device to the EUS FNA needle, the investigators can accurately measure the instant scalar acceleration of the "to and fro" movements.

The investigators propose a prospective, multicenter, randomized, crossover study on 51 patients with solid pancreatic masses to compare an EUS FNA "fast" sampling technique in which the needle acceleration is higher than 1 g to a "slow" technique where the needle acceleration is lower than 1g.

The primary objective of the study is to compare the tissue acquisition rates and the histological diagnosis accuracy between the 2 methods "fast" and "slow". The secondary objectives of the study are to compare the cellularity and quality scores of the obtained specimens between the 2 methods. Another secondary objective is to find a linear relationship between the needle acceleration and the EUS FNA yield (histological diagnosis, sample cellularity and adequacy).

Research questions and aims

The investigators hypothesize that during a EUS FNA (endoscopic ultrasound fine needle aspiration) procedure, the acceleration of the needle that travels through tissue during the "to and fro" movements influence the diagnostic yield.

The investigators aim to test this hypothesis in EUS FNA of solid pancreatic masses by accurately measuring and comparing the diagnostic yield of 2 different needle accelerations, "fast" - where the mean needle scalar acceleration is higher and "slow" - where the mean needle scalar acceleration is lower.

Background and rationale

EUS FNA is an established and recommended technique for diagnostic of solid pancreatic masses. The median EUS FNA sensitivity, specificity, negative predictive value (NPV) and accuracy values for differentiating benign versus malignant lesions are 83% (54-95%), 100% (71-100%), 72% (16-92%) and 88% (65-96%) respectively. The yield of the procedure depends on the operator, lesion and technique related factors:

• operator. A minimum of 25 supervised pancreatic EUS FNA is recommended at the beginning, the diagnostic accuracy plateau improving with experience;
• sedation. General anesthesia significantly increases the diagnostic yield;
• lesion type and location. Lower sensitivities are expected when the punctured mass is found in a pancreas with chronic pancreatitis criteria versus a normal appearing pancreas. Also, lower sensitivity is expected when EUS FNA is performed trans-duodenal (pancreatic head and uncinate process) versus trans-gastric approach (body and tail lesions).
• elastography or contrast enhanced imaging. These may improve the EUS FNA diagnostic yield either by estimating the mass etiology in relation with its elasticity or by finely tuning the exact spot where a solid lesion should be punctured;
• needle diameter. 25 gauge needles may confer an advantage in adequacy relative to 22 gauge needles but confer no advantage with respect to accuracy, number of passes, or complications. 19 gauge needles do not confer an advantage in this moment;
• the presence of stylet only increases bloodiness of the harvested sample;
• the use of suction. The presence of a syringe for aspiration does not appear to increase the diagnostic yield, only increases the bloodiness of the sample. The slow withdrawal of the stylet (the capillary technique) yields better results when using 25 gauge needles, but not for 22 gauge needles. The wet suction technique yields a significant better specimen for 22 gauge needles, but the clinical relevance is questioned;
• the sampling technique. The fanning technique significantly increases the first pass diagnosis rate from 57.7% to 85.7%;
• the number of "to and fro" movements per pass. The optimal number of "to and fro" movements (jabs) is said to be 15;
• the presence of a cytopathologist in the procedure room - rapid on site evaluation (ROSE) increases the yield of EUS FNA by 15 to 20%;
• the minimum recommended number of passes without ROSE is 5 to 7. It seems that 7 passes are non-inferior to ROSE and may be more cost effective;
• finally, the use of fine needle biopsy (FNB) with second generation needles by Cook Endoscopy and by Medtronic Corporation. The debate today is whether this can replace EUS FNA with ROSE;

Recently, the authors of the paper "Multicenter, prospective, crossover trial comparing the door-knocking method with the conventional method for EUS-FNA of solid pancreatic masses" published in "Gastrointestinal Endoscopy" in 2016 hypothesized that that "needle speed" might affect the result of EUS FNA procedure of a pancreatic mass. They designed a prospective study of EUS FNA for pancreatic solid masses using 22 gauge needles, comparing two acquisition techniques: the conventional technique and the "door knocking method". In the later technique, the needle is rapidly pushed towards the stopper, taking care not to exit the outer border of the lesion. The "door knocking method" is a sudden needle deceleration of an unknown value in m/s2.

Overall, the accuracies of the two methods (conventional and "door knocking method") were similar, the quality rates of the histological specimen for the two methods were also similar, only the number of cells were significantly higher in the "door knocking method" (p = 0.03), with discordant results between trans-gastric and duodenal route.

What is the relationship between the needle "to and fro" movement acceleration value and the yield of EUS FNA for solid pancreatic masses?

Firstly, from oocytes aspiration studies, according to Hagen's-Poiseuille law, the velocity of the aspirated flow within the needle tract increases with needle diameter and decreases with needle length. However, from clinical studies so for, one knows that syringe aspiration only increases bloodiness but not sample cellularity. This is explained probably by the fact that only blood (containing small size particles unbound to the extracellular tissue - red blood cells) has a "flow" behavior within the needle tract. Larger particles bound to extracellular space (pancreatic mass cells) have to be firstly detached from the matrix by cutting them with the needle bevel and then aspirated within the needle tract.

In the previous study, as oocytes are meant to be released from ovary, they are not as "tightly" held in the ovarian tissue and this is why aspiration is important. However, one knows that pancreatic cancer has the highest amount of stroma compared to other solid organs and this is why aspiration does not increase the yield of pancreatic EUS FNA. Only needles which can effectively cut and "detach" pancreatic cells may harvest a high cellular sample.

Secondly, from vacuum cleaner physics, particle aspiration is best when the tube does not move very fast on the floor, so as the particles have time to be aspirated.

Applying these two concepts, the investigators hypothesize that the "to" movement will have to have a high acceleration so as to cut, detach and aspire the tissue cells from their surrounding extracellular space and the "fro" movement has to be slower, so as to give the time to the detached tumor cells to be aspired into the needle tube. The investigators aim to verify this hypothesis in EUS FNA of pancreatic solid masses.

Research plan

This is a prospective, multicenter, randomized, crossover study.

Material and methods

Patients with solid pancreatic masses fulfilling the inclusion and exclusion criteria will be allocated to receive EUS FNA with an Olympus linear scope. Each patient will receive 2 EUS FNA passes, 1 "fast" and 1 "slow", with a 22 gauge EUS needle, with suction syringe, employing the fanning technique, 10 jabs for each pass.

A "fast" pass has an advancing mean acceleration jab ("to" movement) higher than 1 g, while a "slow" pass has an advancing mean acceleration jab of less than 1 g (where 1 g equals 9.8 m/s2). Both movements will have a "slow" "fro" withdrawal movement, from the point of maximum advance into the lesion to the lesion entry site.

For each patient, the passes order with be either done as "fast", "slow" or "slow", "fast" in a randomized fashion. The randomization numbers list will be generated on a dedicated website.

The needle acceleration will be measured with an accelerometer (Pocket-Lab) attached to the aspiration syringe, connected to a phone using Bluetooth. Pocket-Lab is a device awarded the "Yale School of Management Grand Prize and Audience Choice Award" and is an "Intel Education Accelerator Partner". It can function as an accelerometer, a gyroscope, a magnetometer, a barometer and a thermometer, all at once. Recorded data is transmitted via Bluetooth to a phone and can be visualized and stored for further analysis. When the device varies its position, speed, acceleration, pressure or temperature, data is captured and transmitted to the phone. The investigators will take advantage on its accelerometer properties.

The Pocket-Lab device will be attached to the EUS FNA needle. As the needle will be moved up and down, the instant scalar acceleration values will be registered to form a chart visualized and recorded on the phone screen. Data can then be exported as a ".csv" file (CSV, comma separated values), translated on a spreadsheet and analyzed in an Excel Table (mean, median, range).

For each patient, acceleration scalar variation data from the 2 passes will be recorded and analyzed. For each pass, the harvested material will be put in a recipient containing absolute alcohol for cyto-block preparation and ulterior histological analysis. Each patient will finally have 2 clearly numbered recipients with harvested material put in alcohol, 1 for "fast" and 1 for "slow" passes. The needle stopper will be fixed at a calculated distance so as the needle will not exit the lesion.

Histological examination will be performed by experienced pathologists from each participating institution. 1 slide for each of the formalin recipient, containing the most available tissue will be stained with hematoxylin and eosin and examined.

At the end of the study, all slides will be scored for cellularity and quality on a 0 to 3 discrete scale by an independent pathologist. This pathologist will not be involved in the initial histological analysis for each patients and will be blinded to the type of pass ("fast" or "slow").

Sample size

For a type I risk error of less than 5% (alpha = 0.05), a type II risk error less than 20% (power beta = 0.80), a histological accuracy for pancreatic masses of 65% with less than 1 g acceleration "to and fro" hand movement and a predicted 90% accuracy with higher than 1 g acceleration hand "to and fro" movement, the calculated sample size is 51 patients.

Data collection

The research protocol, the case report forms (CRF), the informed consent brochures will be submitted by each center to their local independent ethical committee (IEC) for regulatory approval. The trial will be registered at the ClinicalTrials.gov Data will be collected locally in each center in a CRF, translated in an Excel Spreadsheet and then centralized for data analysis. Data will by anonymized and protected according to national and European Union laws.

For each patient, the investigators will collect the following data:

• Date of signed informed consent

• Name (initials on the CRF), date of birth, sex

• Inclusion and exclusion criteria

• Family and past medical history, comorbidities, current medications, including anti-aggregant and anticoagulant therapies (and their stop date before EUS FNA procedure according to the current guidelines)

• European Cooperative Oncology Group, American Society of Anesthesiologists status

• Biological data: full blood count, coagulation parameters, tumoral markers

• Imaging data: abdominal ultrasound, computer tomography scan, magnetic resonance imaging, endoscopic retrograde cholangio-pancreatography;

• Pancreatic tumor: location (head, body, tail), size (largest diameter in mm), solid or mixed type

• EUS FNA: successful tissue acquisition by EUS FNA (yes or no), site of the puncture (trans-gastric, trans-duodenal), adverse events grades according to American Society of Gastrointestinal Endoscopy guidelines

• Histological diagnosis

  • Pass number (1 and 2 slow/or fast, supplementary passages if any - 3 for additional smears and 4 for the aspirated liquid component of a mixed type tumor)
  • Pass method ("fast" or "slow")
  • Cellularity scale (0 to 3)
  • Quality scale (0 to 3)
  • Cytological diagnosis according to a modified Papanicolaou Society I to VII grades (grade VI is "malignant, without any diagnosis" and grade VII is the definite diagnosis of malignancy such as adenocarcinoma, etc

• Scalar acceleration variables: mean, median, range

• Final diagnosis (based on histological diagnosis surgery, biochemical markers, radiology studies and a minimum of 6 months' follow-up)

• Follow up at 30 days for complications

• Follow up at 180 days and date of last news (death and cause of death if applicable)

Planned statistical analysis

Continuous variables will be presented as mean with their standard deviations and range, while categorical variables will be presented in absolute values and percentages.

Comparisons between groups will be done using the McNemar non-parametric test for 2 related samples and the paired T test for the scalar acceleration means comparison.

The linear relationship between the pass acceleration and the EUS FNA yield (histological diagnosis, sample cellularity and adequacy) will be explored by the receiver operating characteristics (ROC) curve and its corresponding area under the curve (AUC).

Statistical analyses will be performed considering a type I error probability less than 0.05.

Inclusion Criteria:

• a solid pancreatic lesion with diameter larger than 20mm, with or without a cystic component, with unknown histology;
• age above 18 years old;
• signed informed consent;

Exclusion Criteria:

• a solid pancreatic mass with a diameter less than 20mm or with known histology;
• a cystic pancreatic mass, without a solid component;
• coagulation disorder (international normalized ratio above 1.5, activated partial thromboplastin time above 42 seconds, platelet count less than 60000/mmc) or impossibility to stop antiaggregants or anticoagulants according to the European Society of Digestive Endoscopy guidelines;
• European Cooperative Oncology Group status 4;
• American Society of Anesthesiology score higher than 3;
• pregnant women;
• age under 18 years old;
• refusal or impossibility to sign informed consent.

At the end of the study, all slides will be scored for cellularity and quality on a 0 to 3 discrete scale, according to the previous work done by Mukai S et al [12], by an independent pathologist, from an institution not involved in the initial histological analysis, blinded to the type of pass ("fast" or "slow").