7. Adjuvant and Neoadjuvant Nonsurgical Therapy of Pancreatic Cancer – Pancreatic Cancer

Surgical therapy is the only potentially curative treatment option for pancreatic cancer (see Chapter 5).

However, long-term survival after resection alone is still below 20% [272]. Long-term recurrence and distant metastasis play a role in tumor recurrence [520]. Therefore, multimodal therapeutic strategies are useful to potentially improve long-term survival in principle. The prerequisite for adjuvant or additive therapy is curatively intended pancreatic resection (R0/R1). Accurate histopathologic workup according to the criteria described in Chapter 5 should be required. With the availability of more effective chemotherapy protocols, perioperative concepts in locally advanced carcinomas should be evaluated individually.

Background

After R0 resection without the presence of distant metastases, five randomized phase III trials and a meta-analysis have shown an advantage in disease-free survival with adjuvant chemotherapy compared to tumor resection alone [521], [522], [523], [524]. Median survival is up to 24 months with monotherapy with gemcitabine (CONKO-001, ESPAC-3) or 5-FU (ESPAC-3). The 5-year survival was approximately 20-45% (JASPAC, Asian patients only) in three randomized trials (CONKO-001, ESPAC-3, JASPAC-01) after adjuvant chemotherapy with gemcitabine, 5-FU, or S1, whereas it was approximately 10% in the arm with surgery alone (CONKO-001). Side effects are relatively low with monotherapy, but more mucositis occurred with bolus 5-FU than with gemcitabine therapy [521], [522], [523]. Two trials (PRODIGE 24/ACCORD 24, ESPAC-4) used combination chemotherapy [527]. In the ESPAC-4 trial, the combination of gemcitabine and capecitabine achieved an improvement in median survival from 25.5 months to 28 months with an unchanged relapse-free survival compared with gemcitabine monotherapy. In the PRODIGE24/ACCORD 24 trial, combination chemotherapy with oxaliplatin, irinotecan and 5-FU/folinic acid (modified FOLFIRINOX, mFOLFIRINOX) improved median survival from 35.5 months with gemcitabine mono to 54 months and an improvement in tumor-specific survival at 3 years from 51.2% with gemcitabine to 66.2% with mFOLFIRINOX was shown [528].

When possible, patients should be encouraged to participate in clinical therapy trials to further improve prognosis and to determine or optimize the optimal duration and type of adjuvant therapy. The course of patients treated outside of clinical trials should be documented as part of tumor documentation/quality assurance with respect to recurrence, survival, and adverse events [521], [522], [525], [526].

Background

From the available literature on adjuvant therapy after resection of pancreatic cancer, there is no evidence that patients of older age, e.g., beyond 70 years of age, benefit less from adjuvant chemotherapy than younger patients. In the studies conducted, the age ranges were 34-82 years [521], 20-79 years [526], and 30-81 years [528].

However, an age restriction of less than 80 years applied to adjuvant therapy with mFOLFIRINOX in PRODIGE 24/ACCORD 24. No data are available on adjuvant therapy with mFOLFIRINOX for patients older than 80 years.

Background

Published trials of adjuvant chemotherapy have included patients with an ECOG performance status of 0-2 or a Karnofsky index of 50-100% or a life expectancy greater than 3 months [530], [531], [532], [533], [534]. Although data show that survival of patients with pancreatic cancer correlates with ECOG [535], there are no data to suggest that patients with an ECOG Performance Status of 2 do not benefit from adjuvant chemotherapy. Because only patients with an ECOG 0-1 were eligible for inclusion in PRODIDGE 24/ACCORD 24, no data on mFOLFIRINOX are available for patients with an ECOG Performance Status of 2 [528].

For patients with an ECOG Performance Status of >2, the indication for adjuvant chemotherapy should be made individually, also taking into account life expectancy [530], [531], [532], [533], [534].

Background

In the CONKO-001 study, R1-resected patients were treated with gemcitabine for 6 months, analogous to R0-resected patients, and thus achieved a statistically significant improvement in survival in a subgroup analysis [521]. Reliable evidence for a longer duration of therapy is currently not available, but a longer additive therapy duration could be useful in terms of tumor biology due to the residual tumor cells Further randomized trials are needed.

Also in the PRODIGE 24/ACCORD 24 trial, the 211 R1-resected patients had a significant benefit from six months of therapy with FOLFIRINOX: HR 0.52; 95%  CI 0.37-0.72 [528].

Data from ESPAC-3 [523] were not included due to the use of the Royal College of Pathologists R classification rather than UICC (R1 defined as tumor cells < 1 mm from the definitive resection margin). For the same reason, this also applies to the data from the ESPAC-4 trial, in which the benefit of adjuvant combination chemotherapy with gemcitabine and capecitabine was limited to R0 resected patients [527].

The CONKO-006 trial, a randomized phase 2 study, included only patients after R1 resection of pancreatic cancer. All patients received additive chemotherapy with gemcitabine for 12 cycles. Supplemental administration of the multikinase inhibitor sorafenib did not improve recurrence-free or overall survival. In a landmark analysis of patients with overall survival of at least 12 months (n=76), there was a significant improvement in overall survival for those who were able to receive 7-12 cycles of additive therapy compared with those who received a maximum of 6 cycles [536] with 29.1 months vs. 18.0 months (HR 0.55; 95% CI 0.32-0.96) [537].

Background

Combination chemotherapy with mFOLFIRINOX leads to a significant and clinically highly relevant improvement in relapse-free survival and overall survival as well as 3-year survival compared with gemcitabine monotherapy. Thus, this is the treatment of choice in patients with a good general condition (ECOG 0-1) under 80 years of age. A reduced dosage of components of the FOLFIRINOX protocol compared to the therapy of metastatic pancreatic cancer as a so-called modified (m)FOLFIRINOX with irinotecan 150 mg/m², oxaliplatin 85 mg/m², folinic acid 500 mg/m² and 5-FU 2400 mg/m²/46h should be considered. However, based on the available evidence but also on clinical experience, this therapy is appropriate only for patients with an ECOG Performance Status of 0-1. To allow all patients to receive the additional benefit of this highly effective therapy, a convalescence of up to 12 weeks after surgery is possible in case of an ECOG > 1 due to surgery/surgical sequelae. These effects were achieved with a dose intensity of approximately 50% mFOLFIRINOX, so that appropriate required dose reductions are possible and reasonable [528].

Background

Adjuvant chemotherapy with gemcitabine remains the standard of care, based on data from the CONKO-001 trial, for those patients for whom intensive combination chemotherapy is not an option, due to an ECOG >1 (i.e., the patient can no longer perform light physical work or work in a sedentary position such as office work) or due to comorbidities [529]. This is also recommended in the current NICE Guidelines [128]. Adjuvant/additive chemotherapy with gemcitabine for 6 cycles resulted in an improvement in 5-year survival from 10.4% to 20.7% and 10-year survival from 7.7 to 12.2 months, respectively, at a median follow-up of 136 months, compared with follow-up alone, in addition to a significant improvement in recurrence-free survival from 6.7 months to 13.4 months. The subsequent ESPAC 3 trial compared 5-FU and gemcitabine as adjuvant chemotherapy after resection of pancreatic ductal adenocarcinoma in a phase III design [525]. This showed that gemcitabine and 5-FU, applied according to the Mayo Clinic bolus protocol, have equivalent efficacy. Both agents significantly improve overall patient survival and equally significantly increase the proportion of patients alive 5 years after resection of pancreatic cancer. However, the toxicity of gemcitabine was lower compared to the 5-FU bolus protocol following the Mayo regimen: grade III/IV mucosal toxicities were significantly more frequent under 5-FU: stomatitis (10% vs. 0%) and diarrhea (13% vs. 2%). Gemcitabine more frequently induced grade III/IV thrombocytopenia (1.5% vs. 0%) [521], [522], [523], [525]. Due to better tolerability and uncomplicated administration, therapy with gemcitabine is to be favored when monotherapy is indicated and is standard in Germany. The use of 5-FU as intravenous bolus administration is now no longer considered the therapeutic standard for the use of 5-FU.

The ESPAC 4 trial evaluated the efficacy of adjuvant chemotherapy with gemcitabine and capecitabine compared with gemcitabine monotherapy in patients with an ECOG of 0-2 [527]. The combination resulted in a prolongation of median overall survival (25.5 vs. 28 months) without an improvement in relapse-free survival. Formally, this is a positive study, as the primary study endpoint "improvement in overall survival" was achieved. However, the lack of improvement in recurrence-free survival should be critically evaluated, especially since in the subgroup analysis primarily patients after R0 resection benefited.

Data for the combination nab-paclitaxel and gemcitabine were collected in the APACT trial. Results to date showed no difference in the primary study endpoint [538], disease-free survival (DFS) assessed independently (i.e., not by a study center/investigator) in the intention to treat (ITT) population at 18.8 (gem), versus 19.4 months (gem plus nab-pac); HR 0.88; 95% CI 0.729-1.063, although this study endpoint makes comparability with other studies difficult. An investigator assessed-DFS was chosen to assess DFS in all studies to date. Assessment of DFS based on external imaging analysis alone is difficult because clinical aspects of recurrence are not considered or are difficult to assess by imaging, such as in the case of peritoneal carcinomatosis . For the commonly used investigator-assessed DFS, there was an improvement by the combination from 13.7 to 16.6 months as well as in OS from 36.2 to 40.5 months. To date, these data are available only a congressional abstrac [539].

Tumor-dependent risk factors such as grading or N-category do not play a role in the decision for adjuvant chemotherapy. The current data do not allow further differentiation of adjuvant therapy.

Background

In some cases, initiation of adjuvant therapy is delayed by perioperative complications or delayed postoperative convalescence. A secondary analysis of the ESPAC-3 trial showed that patients with an initiation of therapy within 12 weeks postoperatively had a comparable benefit from adjuvant chemotherapy as patients with an earlier initiation of therapy. The completeness of adjuvant therapy with 6 cycles was shown to be a relevant factor in improving overall survival in this evaluation.

In the PRODIGE 24/ACCORD 24 study, randomization and thus therapy initiation was allowed between 3 and 12 weeks postoperatively. Data on actual therapy initiation within the trial are not available [528].

In the adjuvant CONKO-005 study as well, in which in principle an initiation of adjuvant therapy up to 8 weeks postoperatively was possible, there was no advantage for an earlier start of therapy (< 6 weeks vs. ≥ 6 weeks) [540].

In a retrospective evaluation of 420 stage I-III patients who received adjuvant therapy between 2004 and 2013, there was no effect for the timing of therapy initiation. Here, initiation of adjuvant chemotherapy within less than 8 weeks was compared with initiation of therapy between 8 to 16 weeks (median OS 20.2 vs. 19.0 months, p=0.97) [542].

Therefore, from the data to date, initiation of therapy within 12 weeks is warranted if a full 6 months of therapy appears tolerable for the individual patient. The 2018 NICE Guidelines [128] does not commit to a specific time point due to lack of directly comparative randomized data and recommends starting therapy only as soon as all 6 cycles can be completed.

Background

All published randomized phase III trials provided adjuvant therapy to patients for 6 months [521], [522], [523], [525]. The recommendation for duration of adjuvant chemotherapy for 6 months is also found in the current versions of the 2018 NICE Guidelines [128] and the 2017 ASCO Guidelines [543].

No additional data from controlled trials are available on the optimal duration of adjuvant chemotherapy in terms of shortening or lengthening.

In the PRODIGE24/ACCORD24 trial, a dose density of only about 50% was achieved for mFOLFIRINOX compared to a dose density for gemcitabine of over 90% [528]. Individual dose modifications and, if necessary, also an adjustment of the duration of therapy must be made depending on the respective general condition of the patient as well as on side effects.

Background

There is no firm evidence of benefit from adjuvant radiochemotherapy. Of the randomized controlled trials conducted to date on this question, the first [545] showed a statistically significant survival benefit (median overall survival 20 vs. 11 months, p<0.03). The main criticism of this study is the small number of participants (n=43), which results in very wide confidence intervals. In addition, patients in the treatment group received long-term chemotherapy for 2 years, so the effect of radiotherapy alone cannot be determined with certainty.

Subsequent RCTs did not confirm this positive result [546], [547], [548], [549]: the EORTC study [550] included 218 patients from 29 centers. The problem with this study is that only slightly more than half of the patients had pancreatic carcinomas, while the remaining patients had periampullary carcinomas with a prognosis different from that of pancreatic carcinomas. The study is therefore underpowered for demonstrating a difference in survival for pancreatic cancer patients. No statistically significant improvement in survival was achieved with radiotherapy.

The ESPAC1 trial by Neoptolemos et al. [523] (multicenter, n=289) also showed no improvement but rather a worsening in survival with radiotherapy. The design (2x2 factorial), randomization conditions and technical delivery of radiotherapy are the main criticisms of this study. The delivery of radiotherapy was not qualitatively monitored. An outdated (from today's point of view) split-course technique was chosen for radiotherapy, in which the low total dose of 40 Gy was applied at 2 Gy per fraction with a pause of 2 weeks after 20 Gy. Such regimens are generally no longer used in radiation oncology when the treatment goal is durable local tumor control. It is well known that pronounced tumor cell repopulation occurs during such pauses in therapy, which diminishes the radiotherapy effect. In addition, delays took place in the radiochemotherapy arm. The median time from surgery to initiation of adjuvant therapy was 46 days in the chemotherapy arm and 61 days in the radiochemotherapy arm. 30% of patients had a protocol deviation in terms of total radiation dose applied or did not receive radiotherapy.

Another RCT by van Laethem et al. [551], as a phase II–trial, showed no differences in toxicity and survival when applying combined radiochemotherapy versus chemotherapy alone. The study was not designed to demonstrate a difference in survival. The available cohort studies (including Yeo et al. [556]) are not powered to show certain results because of the study design and conduct. A meta-analysis also shows no benefit of adjuvant radiochemotherapy in pancreatic cancer [524]. Further studies that meet current methodological requirements are needed to clarify this issue. One such trial for patients with resected carcinoma of the pancreatic head has started recruitment (RTOG 0848/EORTC trial, recruitment target: 950 patients). Only the results of the first randomization of this trial between gemcitabine and gemcitabine with erlotinib have been published [557]. The second part of the randomization with or without radiochemotherapy is not yet available as a full publication.

Recent randomized trials of adjuvant radiochemotherapy after resection of pancreatic cancer confirm the safety of this treatment modality with simultaneous 5-FU chemotherapy or with simultaneous gemcitabine [551], [552]. To minimize side effects, a precise, risk-adapted definition of the target volume and a 3D-compliant radiotherapy technique are necessary. Regarding the definition of the target volume in adjuvant radiotherapy of pancreatic head carcinoma, reference is made primarily to the RTOG Consensus Panel Guideline [558] and also to the recommendations of Brunner et al. [555] and Sun et al. [544] for elective lymph node station irradiation, which are consistent with it.

Regarding the benefit of adjuvant radiotherapy alone, only a single prospective case series from 1992 was identified in the systematic search [554], which included 14 patients (3xT1b, 8xT2, 3xT3, 5xN+). No indication for radiotherapy alone can be derived from these data. In this respect, an indication for radiotherapy alone is also not given.

Background

There is no evidence that additive radiochemotherapy has an advantage over additive chemotherapy. Additive radiochemotherapy has been shown in trials to have better outcomes than no therapy [545], [556], [559]. However, the contribution of the radiotherapy component to this outcome cannot be assessed because direct comparisons between radiochemotherapy and chemotherapy are lacking.

Butturini et al. [560] performed a meta-analysis on the influence of resection status (R0 vs. R1) and adjuvant treatment (chemotherapy vs. radiochemotherapy) after a systematic search from available RCTs. With regard to the distinction R0 versus R1, this was a post hoc subgroup analysis with approximately 1/3 R1 patients (therefore assessed as LoE 2-). Only for adjuvant chemotherapy in R0 patients was there a statistically significant reduction in the risk of death (hazard ratio for death at 44 months 0.65; 95% CI 0.53-0.80), but not for R1-operated patients. For patients after R0 and R1 resection, there was significant heterogeneity in the effect of adjuvant chemotherapy. For R1 patients, radiochemotherapy showed a reduced hazard ratio for death, but this was not statistically significant and the analysis was not adequately powered (HR 0.72; 95% CI 0.47-1.10). Therefore, the authors of the article generally recommend chemotherapy alone in analogy to R0 resection. The uncertainty of the results is due to the unplanned subgroup analysis and the small case numbers in R1 patients; further studies are needed.

Recent randomized trials of adjuvant radiochemotherapy after resection of pancreatic cancer confirm the safety of this treatment modality with both simultaneous 5-FU chemotherapy and simultaneous gemcitabine administration [551], [552]. In van Laethem's study, the incidence of isolated local first recurrence after radiochemotherapy was low at 13% [551]. Primarily, patients without rapid distant progression benefit from improved local tumor control. Therefore, in current trials of adjuvant radiochemotherapy, even after R1 resection, this combined therapy component is used only for non-progressed patients after standard chemotherapy (RTOG 0848/EORTC trial). The best timing for additive radiochemotherapy after resection in the setting of high presumed local recurrence risk, either at completion of adjuvant drug therapy or in the setting of isolated local recurrence, is currently unclear.

The individualized use of radiotherapy in the individual patient should take into account, given the current evidence, that if radiochemotherapy is used early in the course of the disease, timing it too early risks distant metastases becoming clinically leading before local recurrence and thus the patient cannot benefit from the local effect of the radiotherapy component in radiochemotherapy. On the other hand, radiotherapy should not be started too late in isolated or dominantly progressive local recurrence, as target volumes and the risk of side effects increase with increasing tumor volume. The indication should be discussed on a case-by-case basis in the context of an interdisciplinary tumor conference.

Background

The few currently available studies on this question show no clear advantage for neoadjuvant therapy over surgery alone in terms of overall survival in pancreatic cancer that is judged to be resectable [561], [562], [564], [575]. The current 2018 NICE Guidelines also recommend neoadjuvant therapy for resectable pancreatic cancer only in the context of clinical trials [128].

In terms of tumor biology, neoadjuvant systemic chemotherapy also appears reasonable in resectable pancreatic cancer due to the very early metastatic competence of this tumor entity [576]. Potential advantages of this approach include early systemic treatment of existing micrometastases, improved tolerability of chemotherapy, risk reduction for intraoperative tumor cell spread, improved tissue oxygenation, in vivo assessment of success of therapy and potentially increased R0 resection rates. Evidence suggests that a pathologically complete response to neoadjuvant therapy significantly correlates with prognosis [569], [577], [578]. Evaluations of a large meta-analysis with 38 analyzed trials and 3484 patients found an improvement in survival with neoadjuvant therapy for resectable and borderline resectable pancreatic cancer [579] with 18.8 vs. 14.8 months and a significantly stronger effect for those actually resected with 21.6 vs. 15.0 months [580]. These data are supported by a propensity score matched analysis on more than 8000 patients: this showed a median overall survival of 26 vs. 21 months (log rank p<0.01; HR 0.72; 95% CI 0.68-0.78) [580].

Data from the randomized phase III PREOPANC trial showed significant improvement in recurrence-free survival for resectable and borderline-resectable pancreatic cancer with neoadjuvant radiochemotherapy with gemcitabine (300 mg/m² weekly during radiation) compared with adjuvant chemotherapy with gemcitabine alone [566]. In this study, borderline resectability was defined by arterial remodeling of up to a maximum of 90°, in deviation from the criteria mentioned in Section 6.3.1. An improvement in overall survival was seen only for the 120 of the 246 randomized patients who actually underwent resection and were able to start adjuvant chemotherapy at 35.2 versus 19.8 months (p=0.029). The R0 resection rate was significantly increased by preoperative radiochemotherapy at 71% (51 of 72 patients) versus 40% (37 of 92 patients) (p<0.001) [566].

The Japanese JSAP-05 trial enrolled 364 patients with primary resectable pancreatic cancer and either pretreated with neoadjuvant chemotherapy of gemcitabine+S1 or resected directly. This randomized phase III trial showed a significant improvement in median overall survival at 36.6 months versus 26.6 months (HR 0.72; p=0.015) [581]. The results of this study are available only as a congressional abstract to date.

A multicenter phase II trial randomized between immediate resection and neoadjuvant radiochemotherapy each with adjuvant chemotherapy in patients with resectable pancreatic head carcinoma, primary endpoint was overall survival [582]. Of 254 patients to be recruited, 73 could be included and the study was discontinued because of slow recruitment. There was no difference in survival in the intention-to-treat analysis. In the per-protocol analysis with resection, overall survival was 19 vs. 25 months (p=0.79) comparing immediate surgery versus neoadjuvant therapy. Postoperative complications were comparable in both groups [582].

Background

Data from the meta-analysis referenced in Recommendation 7.12 (38 studies analyzed and 3484 patients) demonstrate an improvement in survival with neoadjuvant therapy [579] of 18.8 vs. 14.8 months for resectable and borderline resectable pancreatic cancer,  with a significantly stronger effect for patients actually resected (21.6 vs. 15.0 months). In the PREOPANC trial, also mentioned above, borderline–resectability was defined by arterial vessel walling up to 90°, which differs from the criteria used in these Guidelines (but also internationally). In addition, resectable patients were included in the PREOPANC trial. Results of randomized phase III trials on the value of neoadjuvant therapy in borderline resectable pancreatic cancer are thus still not available. 

As described above, the definition of borderline resectability is problematic in the existing literature: it is often inconsistent in the literature and in almost all studies it is purely anatomical (venous/arterial vessel involvement based on morphology seen in imaging) i.e. without consideration of tumor biology and patient-specific characteristics. In 2017, the International Association of Pancreatology (IAP) established a consensus recommendation that defines additional criteria for borderline resectability in addition to the anatomic ones already summarized in Table 15: these include reduced general status with an ECOG ≥ 2 and an elevated CA19-9 (> 500 IU/ml) or regional lymph node metastasis (confirmed histologically or by PET-CT) [271]. Patients with a preoperative CA19-9 > 500 IU/ml showed a resection rate of less than 70% and a median survival of less than 20 months. Patients with an ECOG ≥ 2 and/or a CA19-9 > 500 U/ml should preferentially receive primary/definitive chemotherapy. Accordingly, a patient who is judged to be resectable based on anatomy (category A) may be classified as borderline resectable by the other two categories (biology, (category B) and general condition (category C)). This may result in a decision for neoadjuvant therapy. Since the consideration of all three categories can be prognostically relevant, a determination of resectability in future studies should no longer be based on anatomical criteria alone [587]. (see also Table 5 on ABC criteria of resectability according to the International Association of Pancreatology (IAP) consensus [271]).

In ASCO LL 2016, preoperative therapy is recommended for the following parameters:

• Suspected extrapancreatic manifestation,
• General condition that does not allow the required abdominal surgical intervention,
• mesenteric vascular infiltration, and
• elevation of CA19-9 (indicative of metastasis) [303].

Data from the ESPAC-5F trial in borderline resectable pancreatic cancer were presented at ASCO 2020, and the results are available as a only congressional abstract to date. This 4-arm phase 2 trial enrolled 90 patients. Direct resection was compared with neoadjuvant therapy consisting of gemcitabine plus capecitabine or FOLFIRINOX or radiochemotherapy [588]. There was no statistically significant difference in resection rates with direct surgery versus neoadjuvant pretreatment (62% vs. 55 %, p=0.668) or R0 resection rates (15% vs. 23%, p=0.721). 44 patients (79%) were able to complete planned neoadjuvant therapy. 1-year survival was improved by neoadjuvant therapy: 40% vs. 77% (HR 0.27; 95% CI 0.13-0.55; p<0.001), however, ESPAC 5F is a feasibility study only.

Background

At the onset of therapy in locally advanced, unresectable pancreatic cancer, a potentially secondary resectable or purely palliative therapy situation can usually not be conclusively assessed. In principle, resectability of the tumor should be pursued as an important treatment goal, since resected patients - as far as can be assessed on the basis of currently available data - have a longer median survival. It should be noted that in addition to the selection bias (predominantly retrospectively collected data), the resectability criteria and the histopathological workup of the resectate are not uniform in the studies or are often not mentioned.

Patients with a sufficiently good general condition should first be given tumor therapy with drugs. The combination chemotherapies FOLFIRINOX [596] and gemcitabine+ nab-paclitaxel [597] each showed tumor response of approximately 30% compared with max. 10% with gemcitabine monotherapy in first-line therapy of metastatic pancreatic cancer and are thus available in principle as effective therapeutic options for locally advanced pancreatic cancer. For patients with an ECOG of 2, who are thus not primarily eligible for first-line therapy with FOLFIRINOX, gemcitabine + nab-paclitaxel can be administered in a dose-reduced form [598].

Background

The literature shows conflicting results on the effects of IORT in terms of improved survival and in terms of potential complications for patients with resectable pancreatic cancer [599], [600], [601], [602], [603]. Due to the retrospective nature of the studies, confounders and bias cannot be excluded; in addition, patients were often analyzed over a long period of time with changing surgical standards. The level of evidence for these retrospective cohort studies was formally assessed as 2b in each case [599], [600], [601], [602], [603]. Due to long recruitment periods, low case numbers and retrospective surveys, a 2b- was assigned in each case.

For patients with pancreatic cancer judged to be unresectable, the literature also does not yield a clear benefit with regard to IORT, since, first, the comparison of postoperative percutaneous radiotherapy vs. IORT is not available and, second, a statistically significant improvement in survival could also be shown with chemotherapy without radiotherapy [604], [605], [607], [606]. The cohort studies were also rated 2b. In addition, very few data on side effects are available.

The use of IORT outside of prospective, controlled, desirably randomized trials cannot be recommended based on the available literature.

Background

Due to insufficient studies, no recommendation can be made on this question at present.

Background

Both FOLFIRINOX and gemcitabine+nab-paclitaxel achieve good results in the therapy of metastatic PDAC in terms of tumor response with respect to achieve partial remission. Several retrospective evaluations demonstrate feasibility and consistently report benefit from neoadjuvant chemotherapy [593], [594]. Neither regimen can be clearly favored because no fully published data from randomized trials are available for a direct comparison. In a synopsis of all available data from retrospective evaluations only, there seems to be a better efficacy for neoadjuvant chemotherapy with FOLFIRINOX. Complementary radiation does not improve secondary resectability and/or survival [608].

A meta-analysis of 13 trials involving 689 patients, 355 of whom had locally advanced pancreatic cancer, revealed very good efficacy and safety for therapy with FOLFIRINOX [608]. The analysis showed a median overall survival of 24.2 months and a median progression-free survival of 15.0 months for combination therapy. The rate of secondary pancreatic resection was 28% [608].

The NEOLAP study [609] supports data on the efficacy of induction chemotherapy with gemcitabine+nab-paclitaxel (alone for 4 cycles or in combination for 2 cycles plus 4x FOLFIRINOX) in these patients. Of 168 patients included, approximately 1/3 underwent secondary resection. Resected patients showed a median survival of 27.4 months compared to that of non-resected patients at 14.2 months (p=0.0035).

To date, locoregional intensification of treatment with radiochemotherapy has been randomized only in the phase III LAP-07 trial. Here, 442 randomized patients after 4 months of chemotherapy with gemcitabine (+/erlotinib) showed no difference in median overall survival with 16.5 months after chemotherapy alone versus 15.2 months after adjuvant radiochemotherapy. (HR 1.03; p=0.83). Re-induction of chemotherapy was significantly delayed to 6.1 months versus 3.7 months after radiochemotherapy (p=0.02) [589]. Thus, the implementation of radiochemotherapy can be considered in patients for whom a break in (chemo) therapy is important and reasonable. Importantly, the LAP trial did not include exploration for secondary resection. One trial that did (CONKO-007) has been completed, but the results are not yet available. It should be noted that the induction chemotherapy used, gemcitabine (+/- erlotinib), can now no longer be considered standard chemotherapy. FOLFIRINOX, gemcitabine+nab-paclitaxel or gemcitabine monotherapy are recommended treatment options here - depending on the patients ECOG status.

The duration of intensive systemic (induction) therapy in pancreatic cancer initially judged to be borderline resectable or locally advanced has not been clearly established. In a retrospective evaluation of 123 patients with PDAC judged to be borderline resectable and 71 patients with PDAC judged to be locally advanced, the following factors were shown to be prognostically relevant: a duration of chemotherapy of at least 6 cycles, the decrease in CA19-9 with chemotherapy and the presence of a major pathological response [309].

Background

After neoadjuvant therapy, resectability of borderline resectable (BR) or locally advanced pancreatic cancer (LAPC) cannot be reliably evaluated with currently available imaging (CT, MRI, PET-CT): even if there is no clear morphological response seen in imaging, surgical exploration sometimes reveals only residual scar tissue without vital tumor at the critical arterial structures [615]. In retrospective studies, secondary resection rates of 30-60% have been achieved, especially after FOLFIRINOX, with R0 resection possible in up to 90% of cases [311][593]. 

Data from two retrospective evaluations of 110 and 223 patients with BR and LAPC, respectively, confirm the drop in CA19-9 as an important prognostic marker prior to secondary planned resection with only very limited power of cross-sectional imaging [611]. In addition, a neoadjuvant therapy duration of 6-8 months seems to be prognostically beneficial [311][616]. Here, after initial chemotherapy, imaging showed partial remission in 48% of patients, stable disease in 47%, and disease progression in 5%. In the 61 patients that underwent secondary resection, the extent of remission had no impact on overall survival [312]. In another retrospective case series of 129 patients with BR and initial chemotherapy, imaging ultimately showed conversion to clearly resectable tumor in only one patient (0.8%). 69% of patients had stable disease on imaging, 12% had partial remission, and 19% had progressive disease. Nevertheless, 66% of patients (of which 80% were R0) could be resected. Thus, a median overall survival of 30 months was achieved. A limiting factor for surgical exploration after neoadjuvant therapy is a persisting distal venous vessel involvement with extension below the duodenal inferior border [271], making venous reconstruction technically impossible. In this case, surgical exploration can be omitted.