Needle-Tract Seeding of Pancreatic Cancer after EUS-FNA: A Systematic Review of Case Reports and Discussion of Management

Simple Summary Needle-tract seeding (NTS) is a rare but serious complication of Endoscopic Ultrasound-guided biopsy of pancreatic adenocarcinoma. Due to its very low incidence, there is lack of evidence about its management. The aim of our systematic review is to deepen the knowledge about this entity and highlight therapeutic approaches. After a systematic search we retrieved 45 cases, plus one from our center. The majority of patients (87.1%) underwent surgical resection of the primary pancreatic tumor, of which only 55.9% received neoadjuvant and/or adjuvant chemotherapy. NTS occurred at a median of 19 months after primary diagnosis, with a secondary surgical approach in 82.4% of patients and a median overall survival of 26.5 months compared to 15.5 months when treated with chemo/radiotherapy. NTS is rare, generally occurs late and might be treated aggressively with good results. As only a low rate of patients developing NTS underwent (neo)adjuvant chemotherapy, one may speculate that it might be protective. Abstract Needle-tract seeding (NTS) has been sporadically reported as complication of Endoscopic Ultrasound (EUS)-guided aspiration (FNA) in pancreatic adenocarcinoma (PDAC). However, the evidence of its treatment and outcome is sparse. Adhering to PRISMA guidelines, we conducted a systematic review of EUS-FNA NTS cases of PDAC and analyzed their management and outcome. Up to September 2022, the search query retrieved forty-five cases plus an unpublished case from our center, for a total of forty-six; 43.6% were male, with a mean age of 68.6 years. Thirty-four patients (87.1%) underwent an initial surgical resection, with only 44.1% and 5.9% undergoing adjuvant and neoadjuvant chemotherapy, respectively, and 5.9% undergoing both. The NTS nodule was mostly located in the posterior gastric wall, developing at a median of 19 months after primary resection; 82.4% underwent surgical resection of the seeding, while for 17.6%, palliative chemotherapy treatment. Follow-up after NTS diagnosis and treatment was reported for only twenty-three patients: when NTS was treated with surgery, the median overall survival was 26.5 months compared to 15.5 if treated with radio/chemotherapy. NTS after EUS-FNA of PDAC occurs late and might be treated aggressively with good results. Interestingly, only a low number of patients developing NTS underwent chemotherapy for the primary cancer, suggesting its possible protective role.

Two authors (LA and RPDLP) screened all titles and abstracts for eligible articles based on predefined eligibility criteria. Full-text manuscripts were evaluated, and references of the identified manuscripts were examined for further potentially relevant articles not identified by the initial search.
This review has been registered in PROSPERO-National Institute for Health Research, ID -CRD42022356314.

Eligibility Criteria
Inclusion criteria were: (a) case reports or case series concerning patients with a pancreatic lesion undergoing EUS with FNA/FNB; (b) development of a lesion that is highly suspect or cyto/histologically confirmed as deriving from the primary pancreatic tumor; (c) the lesion is described as being located in the trajectory of the previous FNA puncture site, and therefore, highly suggestive for NTS. Exclusion criteria were: (a) pancreatic lesions different from PDAC; and (b) peritoneal recurrences.
In the event of duplicate publications, the most recent or more complete one was selected.

Data Extraction
The following data were retrieved and registered in a dedicated spreadsheet: country of origin, publication year, patient gender and age, PDAC-EUS and surgical specimen size, site, caliber and type of needle used to perform FNA on the primary PDAC, number of FNA passes, puncture route, TNM stage and management of the primary PDAC, histological diagnosis, grading and R status of the resection, localization of seeding, time lapse between EUS-FNA of the primary tumor and diagnosis of seeding nodule, seeding nodule management, and patient outcome. Whenever any of this information was not provided in the manuscript, we attempted to retrieve them by contacting the corresponding author by email.
A case of NTS diagnosed at our center was also included.

Case Report
We report the case of a 57-year-old male, referred to our center for the follow-up of alcoholic chronic pancreatitis, who underwent pancreatico-jejunostomy in 2015 for recurrent acute flares. In March 2016, at a follow-up EUS, a 20 mm mass of the pancreatic tail was diagnosed. An EUS-FNA of the lesion with a 25-gauge Menghini-tip needle (Expect SlimLine, Boston Scientific, MA, USA) was carried out through the posterior gastric wall (2 passes) with a slow-pull technique, and the cytological examination revealed PDAC. The patient had a Karnofsky performance status score of up to 80%, and the pre-operative computed tomography (CT) scan confirmed the 20 × 20 mm lesion of the tail, invading the splenic vessels; levels of Ca19.9 were slightly elevated (175 U/ml). In April 2016, the patient underwent upfront distal pancreatectomy with splenectomy for T2N0M0 R0 G3 (TNM 8th edition) PDAC, with 14 negative lymph nodes removed. Follow-up management included adjuvant chemotherapy with 6 cycles of gemcitabine and radiological surveillance with CT scan every 3 months.
A CT scan performed 9 months after surgical resection showed the onset of a focal mass in the posterior gastric wall. EUS revealed a 25 mm hypoechogenic mass arising from the muscularis propria layer of the posterior gastric wall, stiff at elastography, isoenhanced after the intravenous injection of SonoVue ® (Bracco, Milan, Italy) contrast agent with a hypoenhanced core ( Figure 1C). Initial differential diagnosis was between a malignant, although small, gastrointestinal stromal tumor (GIST) and a case of seeding due to the previous EUS-FNA. The cytological examination of the EUS-FNA tissue acquired reported a localization of PDAC. acquired reported a localization of PDAC.
The patient underwent atypical gastric resection in April 2017 with a histological confirmation of a 35 mm gastric recurrence of the previous PDAC from the submucosal layer to beyond the serosa. The mass was located in the site of the previous EUS-FNA puncture, strongly suggesting a EUS-FNA NTS.
He subsequently underwent further adjuvant chemotherapy, with evidence of liver metastases onset at 6-month CT scan, and patient exitus on April 2018, 25 months after the first diagnosis of PDAC.

Selection of Included Studies
The search initially retrieved 1374 results from Pubmed and 326 results from Scopus. After a screening of titles and abstracts, 30 reports (either case reports or series) were selected from Pubmed with no additional reports from Scopus. Another 3 reports were identified from hand-searching of the references of the selected manuscripts and included, for a total of 33 reports [11,17,19,. (Figure 2). The patient underwent atypical gastric resection in April 2017 with a histological confirmation of a 35 mm gastric recurrence of the previous PDAC from the submucosal layer to beyond the serosa. The mass was located in the site of the previous EUS-FNA puncture, strongly suggesting a EUS-FNA NTS.
He subsequently underwent further adjuvant chemotherapy, with evidence of liver metastases onset at 6-month CT scan, and patient exitus on April 2018, 25 months after the first diagnosis of PDAC.

Selection of Included Studies
The search initially retrieved 1374 results from Pubmed and 326 results from Scopus. After a screening of titles and abstracts, 30 reports (either case reports or series) were selected from Pubmed with no additional reports from Scopus. Another 3 reports were identified from hand-searching of the references of the selected manuscripts and included, for a total of 33 reports [11,17,19,. ( Figure 2).
As there were some missing data in all the retrieved studies, we tried to retrieve them by contacting all corresponding authors of these 33 papers, but we only managed to receive a reply from three of them [19,44,45], although the requested data were available for only one of them. We did not include a recent Japanese nationwide survey [48] investigating the status of NTS after the EUS-FNA of primary pancreatic tumors in surgically resected patients because it did not satisfy the eligibility criteria; however, we analyzed its results in the discussion.

Features of the Included NTS Cases
The unpublished newly developed case from our center was added to the 45 literature cases from the 33 reports for a total of 46 cases, described between 2005 and 2022 [11,17,19,.
Seventeen of the 46 patients (43.6%) were male, with a mean age of 68.6 years. Fortythree cases (93.5%) arose from the pancreatic body/tail, 3 (6.5%) from the pancreatic head; all 46 cases (100%) were histologically diagnosed as Pancreatic Ductal Adenocarcinoma (PDAC). As there were some missing data in all the retrieved studies, we tried to retrieve them by contacting all corresponding authors of these 33 papers, but we only managed to receive a reply from three of them [19,44,45], although the requested data were available for only one of them. We did not include a recent Japanese nationwide survey [48] investigating the status of NTS after the EUS-FNA of primary pancreatic tumors in surgically resected patients because it did not satisfy the eligibility criteria; however, we analyzed its results in the discussion.

Features of the Included NTS Cases
The unpublished newly developed case from our center was added to the 45 literature cases from the 33 reports for a total of 46 cases, described between 2005 and 2022 [11,17,19,.
The caliber of the needles used to acquire tissue for the primary cancer diagnosis was 22 G in 27 cases (64.3% of cases), but other needles of 19, 20, 21, or 25 G were also employed; in terms of number of needle passes, these were >1 in 33 (94.3%) of the reported cases and varied between 1 and 5. Puncture route was trans-gastric in 44 cases (95.6%).
Thirty-four patients (87.1%) underwent surgical resection of the primary pancreatic lesion, of which only 2 (5.9%) received neoadjuvant chemotherapy, 15 (44.1%) received adjuvant chemotherapy, and 2 (5.9 %) received both. Five patients (12.8%) were not resected and were treated with primary chemotherapy and/or radiation therapy. Thus, the majority of the reported NTS cases were treated with surgery alone on the primary tumor.
Time between EUS-FNA and the diagnosis of NTS ranged between a few days to a maximum of 42 months, with a median onset after 19 months. In 42 (91.3%) patients, NTS The caliber of the needles used to acquire tissue for the primary cancer diagnosis was 22 G in 27 cases (64.3% of cases), but other needles of 19, 20, 21, or 25 G were also employed; in terms of number of needle passes, these were >1 in 33 (94.3%) of the reported cases and varied between 1 and 5. Puncture route was trans-gastric in 44 cases (95.6%).
Thirty-four patients (87.1%) underwent surgical resection of the primary pancreatic lesion, of which only 2 (5.9%) received neoadjuvant chemotherapy, 15 (44.1%) received adjuvant chemotherapy, and 2 (5.9 %) received both. Five patients (12.8%) were not resected and were treated with primary chemotherapy and/or radiation therapy. Thus, the majority of the reported NTS cases were treated with surgery alone on the primary tumor.
Time between EUS-FNA and the diagnosis of NTS ranged between a few days to a maximum of 42 months, with a median onset after 19 months. In 42 (91.3%) patients, NTS was reported in the gastric wall, whereas 2 (4.4%) were reported in the remaining pancreas itself, 1 (2.15%) in the duodenum, and 1 (2.15%) in the cardia.

Treatment of the Seeding Site and Outcomes
Concerning the management, 28 (82.4%) patients diagnosed with NTS underwent a new surgical resection as treatment for the seeding nodule, mostly localized in the posterior gastric wall, while 6 (17.6%) underwent palliative chemotherapeutic treatment, being, therefore, managed as a metastatic spread of the disease instead. For 12 patients (26.1%), no information about NTS treatment was available, and for 24 patients (52.1%), no information about overall survival was available. The corresponding authors of 20 of these patients were contacted, but replies for only 10 patients were received, with most of them replying that no information could be retrieved except for one case [44].
Of the 28 patients undergoing surgical resection of the gastric NTS, disease re-recurrence after NTS removal was described in 6 patients (1 in the liver, 2 in both liver and peritoneal, 2 peritoneal, and 1 had a second gastric recurrence).
The median overall survival of NTS treated with surgery was 26.5 months compared to the 15.5 months of survival of patients treated with oncological treatment alone. Although a statistical comparison was not considered methodologically sound, an aggressive treatment seemed to be associated with better outcome.
Other information about chemotherapy, surgical treatment, EUS-FNA, seeding nodule, tumor stage, and patient management are reported in Table 1.

Conclusions
NTS is a rare but serious long-term complication of EUS-FNA, with a very low prevalence that has been reported to range between 0.003% and 0.009% [14]. A recent systematic review and meta-analysis by Facciorusso et al. [15] calculated this incidence to be higher, up to 0.4% when focusing on pancreatic adenocarcinoma, albeit without an apparent impact on prognosis, as shown by the similar rate of carcinomatosis onset reported in patients undergoing EUS-FNA or not. In that meta-analysis, nonetheless, the management of NTS and how this could impact the prognosis was investigated.
With the widespread use of EUS with tissue acquisition for pancreatic lesions, NTS localizations may be expected to become more frequent, and their standardized management is therefore advisable. The management of NTS localization is, instead, not standardized and several options, from surgical resection to palliative chemotherapy could be considered.
The present systematic review reported data on 46 cases that fulfilled our stringent inclusion criteria for cases in which PDAC seeding on the needle track was deemed extremely likely. In some of these cases, the authors performed molecular analyses of the NTS tissue and compared it to the primary pancreatic tumor to confirm the etiology of the nodule based on the striking similarity of genomic landscape [37]. We excluded cases presenting with seeding in the peritoneum. In case of carcinomatosis onset, it could indeed be difficult to distinguish between procedure-related seeding and disease metastatic dissemination. In keeping with our approach, the study by Facciorusso et al. proved that the use of EUS-FNA does not seem to increase the rate of peritoneal carcinomatosis [15], and therefore, it is unlikely that cases of NTS with peritoneal carcinomatosis are genuine seeding cases.
We gathered information to homogenize available data from previous sparse reports and hereby present the largest combined analysis on NTS features in 46 patients. Although our aim was to retrieve all of the cases' missing data by contacting the authors, the long time gap between the first and the last reported cases did not allow us to obtain complete information about the clinical course.
Most of the 46 included NTS cases were located in the gastric wall (91.3%), a minority being in the duodenum, in the cardia, and in the pancreas itself. This is not surprising, as most of the seeding cases occurred after trans-gastric biopsies of lesions of the body-tail that would have their stomach preserved when surgically resected [11,17,19,[21][22][23][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42]45,46]. On the other hand, in cases of lesions in the pancreatic head, the eventual onset of seeding could remain undetected, as the wall pierced by FNA route (duodenal bulb or second portion) would be removed during surgery [43], but no microscopic pathological investigation would be performed on such site; second, considering the short overall survival of PDAC patients [1], NTS foci may not have enough time to emerge and become detectable during follow-up investigations.
One may therefore hypothesize that the real incidence of NTS in general may be underestimated and underdiagnosed.
An interesting issue arose regarding the timing of NTS diagnosis: there was a wide range, from few days to 42 months (median 19 months), even in patients with similar disease course and treatments. This heterogeneity is likely explained by the variability in terms of growth rate that is also observed in primary PDAC treated with chemotherapy or experiencing recurrence after radical surgery [13]. In this view, NTS might be considered an interesting model of PDAC cell-growth patterns and timeline and help clarify the pathophysiology of disease recurrence. Indeed, how much time it takes for PDAC to grow from single cells to a solid nodule until symptom onset remains largely uncertain.
In terms of NTS management, it is interesting to highlight how in a minority of the reported cases (17.6%), NTS was considered and, therefore, managed as a metastatic site of disease, usually bearing an expected survival of <12 months and, therefore, receiving either chemotherapy or palliative treatments alone. Most patients (82.4%) were instead treated with secondary surgical excision of the NTS nodule, as their NTS was considered as a de novo T2N0M0 lesion [17]. This more aggressive approach succeeded in eradicating the recurrent disease in some cases. The lack of data about the patients' overall survival and the small number of reported cases did not, however, allow meaningful comparison between different treatments and their outcomes, despite NTS treated with surgery showing a longer median overall survival (26.5 months) compared to those treated with chemotherapy or palliation (15.5 months).
A very recent nationwide survey conducted in Japan [48] reported data about NTS in 12,109 patients who underwent surgical resection of primary pancreatic tumors. In that study, the NTS incidence (40 patients-0.33%) was low and localization (97.4%) was mostly in the gastric wall. Twenty-five patients (65.8%) underwent surgical resection of the seeding lesion, with a longer overall survival compared to those who did not. Most data from that survey seem, therefore, in keeping with the present study. Unfortunately, the lack of information about neoadjuvant or adjuvant chemotherapy in that cohort strongly affects the clinical relevance of the results. Also, as the data was only derived from a survey, with a different methodology and lack of detailed information, we did not include these cases in our qualitative analysis of the literature.
The reported cases of NTS support the concern of EUS being potentially risky for the diagnosis and tissue acquisition of resectable pancreatic neoplasms. Some researchers have indeed highlighted a possible role of EUS-FNA in promoting distant metastasis, analyzing cell-free DNA plasma concentration before and after EUS-FNA of PDAC and showing how EUS-FNA is associated with increased plasma concentration of cell-free DNA and increased detection of mutant KRAS [49]. As a matter of fact, many NTS cases were located in the gastric submucosa, a layer characterized by abundant blood and lymphatic vessels, supporting the pathogenic mechanism of NTS in promoting metastatic spread [44]. Nonetheless, several studies have demonstrated that EUS-FNA of PDAC is associated with neither worse overall nor recurrence-free survival [17][18][19].
Since the first publication of reports about NTS after EUS-FNA, several novel methods and devices have been introduced to reduce the number of required punctures [13]. In the present cohort, 22 G was the most commonly employed needle (64.3%), and almost all patients underwent more than 1 pass (94.3%). Interestingly, it has been advocated to minimize the number of FNA/FNB passes to reduce the risk of complications, but possibly also to reduce the risk of seeding, for example, with the use of rapid onsite evaluation(ROSE) when available, or with the use of FNB needles to acquire more tissue with less passes. Furthermore, an additional suggestion could be to wipe the needle, flushing it with sterile saline, between each pass to remove any remaining tumoral cell. Although there are some reports of confirmed NTS after single passes [17], the majority of published cases reported NTS to occur after multiple passes; considering the low number of studies on the topic, nonetheless, the association between a defined needle or the number of passes and NTS onset as a complication cannot be ascertained. Notably, an additional interesting aspect regarding the FNA technique applied was observed: the "slow-pull" technique applies, in fact, less pressure compared with the "suction" one, possibly leading to an easier cell fall in the needle route and, therefore, a higher risk of seeding. However, as only few of the reports described the technique of the biopsy, a comparative analysis was not possible.
While the event of seeding during EUS for PDAC should be kept in mind, we might speculate that this risk may be reduced by neoadjuvant and adjuvant treatments. In fact, notably, only 55.9% of the patients with reported NTS who underwent surgical resection of the primary pancreatic tumor had undergone neoadjuvant and/or adjuvant chemotherapy for the primary tumor, despite current guidelines recommending a perioperative chemotherapy for all fit patients [50]. One may speculate that isolated tumor cells may be easily eliminated by an active chemotherapy administered either before or soon after surgery, reducing the potential risk of NTS. The wider diffusion of neoadjuvant polychemotherapy before surgery [2] will, therefore, likely reduce NTS incidence. Large studies comparing the rate of NTS in patients treated or not with different perioperative chemotherapy policies (none, adjuvant, neoadjuvant, and both) are needed to confirm this hypothesis.
In conclusion, the present systematic review gathered available data on NTS after EUS-FNA pooling the largest dataset of reported cases to date, with data confirming that this rare complication usually occurs late (median onset after 19 months from EUS-FNA) and should be distinguished from "typical" distant disease recurrence since, when treated more aggressively with repeated surgery, the overall survival seems longer (26.5 months) compared to that of patients treated with palliative treatments (15.5 months).