Basic Science and Translational Studies

Since the early 1990s there has been a rapid increase in in vitro and in vivo animal investigations evaluating the effect of single or combined perioperative interventions on tumorigenesis and metastasis. Most of those preclinical studies indicate that surgical stress, volatile anesthetics, and opioids can mediate cancer progression by (1) acting on signaling mechanisms that promote cell survival, proliferation, epithelial-mesenchymal transition, and metastatic cell behaviors; (2) inducing inflammation and suppression of the immune response against cancer; (3) triggering angiogenesis; and (4) promoting cancer “stemness.” ^, On the other hand, laboratory studies suggest that a reduction of the surgical stress and inflammatory response via regional anesthesia, infusions of the local anesthetic lidocaine, beta blockers, and antiinflammatory agents, such as nonsteroidal antiinflammatory drugs (NSAIDs), reduce the metastatic burden associated with surgery. ^,

Unfortunately, in the context of perioperative cancer medicine, the translation of experimental animal research into humans has largely failed. For instance, while studies in rodents indicate that minimally invasive surgery reduces cancer progression, well-conducted clinical studies have shown no benefits or the opposite effects in terms of benefits in survival. In addition, significant advances in the areas of genomics, proteomics, and metabolomics have permitted clinicians and clinical researchers to reclassify and treat cancers based on their actionable targets instead of histological subtypes of disease. ^, However, it is still unknown how drugs with multiple potential receptor targets or epigenetic effects, such as anesthetics and some analgesics, can affect oncological outcomes in an era of “precision oncology.”

Clinical Endpoints

When considering randomized controlled trials (RCTs) designed to test the efficacy of perioperative interventions in patients with cancer, improvements in overall survival and quality of life should be considered as the two most important endpoints. ^, While other survival endpoints, such as recurrence-free survival, disease-free survival, progression-free survival, or biochemical-free survival are also commonly used in clinical trials, they do not always correlate well with improvement in overall survival. In other words, when intervention A proves to significantly extend recurrence-free survival time in favor of intervention B, the actual impact on overall survival may not be clinically relevant. ^,

Furthermore, it has to be understood that most clinically available biological surrogate endpoints are poor indicators of the ultimate goal in the treatment of cancer, which is cure (long overall survival). In other terms, a single biomarker or a combination of biomarkers that can be easily measured and/or modulated by perioperative interventions is not always highly predictive of overall survival. As an example, findings from randomized controlled studies in breast cancer patients suggesting that the use of paravertebral block anesthesia caused a significant reduction in biomarkers of inflammation and angiogenesis did not translate in a longer overall survival.

Short- and long-term psychophysical and social recovery after surgery could be considered as endpoints of postoperative quality of life. From the patients’ point of view, recovery is defined as return to “normality.” ^, In the patient with cancer, the definition of recovery can be broadened to their ability to continue with the cancer treatments (adjuvant therapies), such as chemotherapy or radiotherapy. For some malignancies, such as breast cancer, early return to adjuvant therapies is associated with benefits in survival. However, it must be considered that a disconnect between clinical indicators of recovery and patients’ perception of recovery can exist before patients are able to return to their cancer journey, thus patients may still receive adjuvant therapies before achieve psychosocial recovery. We strongly believe that in all patients but specially in those with cancer, postoperative quality of life and recovery should be measured using patient-centered multidimensional tools, in addition to traditional and institutional measures of recovery, such as adequate pain controlled, minimum nausea and vomiting, ambulation, and return to basic physiological functions. ^,

In summary, it is of outmost importance to consider overall survival as the main endpoint in studies testing the efficacy of perioperative interventions in cancer outcomes.

Retrospective and Population Database Studies

Recent improvements in data capture, data integration from large registries, and the use of sophisticated statistical models permit clinical and translational investigators to evaluate associations between different perioperative interventions—in the so-called “real clinical practice”—and oncologic outcomes. Thus it is not surprising that more than 30 retrospective studies have been conducted to evaluate the association between the use of different anesthetic or analgesic techniques with improvements in recurrence-free survival or overall survival. While this type of research methodology can be seen as relatively inexpensive with results that are obtained in a relatively short period of time, it can require extensive technical support and personnel dedicated to “clean, mine, and combine” large databases that are not always created with investigational purposes.

Retrospective studies and large registry databases have other disadvantages. First, missing information is a significant source of bias and a confounding factor. In addition, most published studies do not report how missing data were analyzed. In our opinion, the use of multiple imputation analysis is generally accepted as a good strategy to handle missing information since the list-wise deletion or single imputation method can significantly reduce the sample size and or assign values incorrectly to missing variables; both of which are sources of bias. ^, Second, while sophisticated statistical analyses (i.e., propensity score matching) have been created to diminish imbalances in baseline demographic characteristics, the lack of randomization introduces selection bias. ^, Moreover, even when propensity scores analyses are designed to control for nonrandom treatment effects, retrospective studies tend to overestimate treatment effect. Another significant issue arising from the use of large database analysis can be the exaggerated interpretation of the findings based on traditional thresholds for type I error (a = 0.05). Researchers and clinicians demonstrating statistically significant differences not always or necessarily suggest a clinically relevant effect. Therefore, they use complementary approaches, such as the Bonferroni correction to obtain a more “strict” alpha error and thus avoid the presence of false positive findings, or the use of confidence intervals ( estimates of effect ) seems more prudent. In line with this idea, The New England Journal of Medicine has recently adopted new guidelines for statistical reporting, emphasizing the importance of measures of estimates of effect over P -values for primary outcomes. Last and perhaps most important, it is to remember that retrospective studies only evaluate an association and not cause-effect between an exposure variable and the studied outcome or outcomes. Hence findings and conclusions from retrospective studies should be considered hypothesis generating and their findings may require rigorous experimental testing by RCTs.

In summary, researchers will likely continue to use databases or large registries to evaluate associations between perioperative interventions and cancer-related survival or overall survival. However, in our opinion the results from such studies will be complementary to those obtained from RCTs. Last, the clinical significance of the findings generated from large registries and database studies can be limited by the lack of significant external (generalizability) and internal (bias) validity.

Randomized Controlled Trials

Over the last two decades, there has been an increase in the number of RCTs designed to address causation between regional anesthesia, intravenous lidocaine, beta blockers, and cyclooxygenase inhibitors and longer survival. However, the results of only one of those trials was recently reported by Sessler et al. who tested the hypothesis that in comparison to opioid and volatile-based general anesthesia, paravertebral block anesthesia in combination with propofol sedation (or general anesthesia) would reduce cancer recurrence in women with no metastatic breast cancer having mastectomies without extensive breast reconstruction. The authors found no clinically relevant differences in survival between the two treatment groups.

Clinical decisions should be based on high-quality unbiased evidence. In our opinion, RCTs are the best methodology to experimentally compare the efficacy of different perioperative interventions. One of the main premises of any RCT is to reduce any potential selection bias by evenly distributing patient-related variables across treatment groups and assuming that prognostic factors of cancer progression do not interact with the intervention’s (anesthetics or analgesic) specific effect. Taking into consideration our premise for RCTs as the standard for testing causation (cause-effect relationship) between anesthetic techniques and cancer outcomes, we should therefore assume that patients enrolled in comparative studies testing a perioperative intervention A versus B should have at the completion of the study similar demographics, tumor-related variables, neo- or adjuvant treatments, and neither of them will interfere with the potential effects of the allocated anesthesia or analgesia technique.

Large RCTs in perioperative medicine require extensive funding to institute a robust clinical research infrastructure and to support costs involved with screening and necessary diagnostic tests, follow-up programs for monitoring and evaluating needed interventions, data management, and research personnel. Underfunded clinical research programs frequently result in RCTs that are not adequately powered for meaningful and clinically impactful endpoints. Another limitation is the lack of properly trained personnel dedicated to clinical research in perioperative medicine. Given the complexity of the perioperative period, and the multitude of factors that can affect clinical outcomes, poorly designed and conducted RCTs can lead to erroneous findings or conclusions. Poorly constructed RCTs are frequently limited by significant bias, and methodological flaws such as a lack of blinding or allocation concealment, lack of statistical power, and significant failures in follow-up.

One of the major challenges of conducting and completing large RCTs in the areas of cancer and perioperative medicine research is related to difficulties in enrolling patients into studies. In general, less than 10% of subjects with cancer are recruited in clinical trials, which translates to several months or even years of patients’ enrolment and study completion. For instance, the results of a major RCT testing the efficacy of paravertebral block anesthesia to reduce breast cancer recurrence were released almost a decade after the study initiation. Since new cancer treatment strategies can be discovered during follow-up periods, the translational value of those studies may be hampered. Thus after the completion of such RCTs the external generalizability of their results can be questioned. Alternatively, pragmatic clinical trials conducted in real-world settings, in which study drug(s) or intervention(s) are tested against usual care (rather than placebo), have been proposed to increase patients’ eligibility and enrolment, enhance the generalizability of the results, and reduce costs. ^,

Conclusion

The field of perioperative oncology medicine is continuously evolving due to important strides in basic, translational, and clinical research. On the one hand, experimental laboratory research will help to answer mechanistic questions on how perioperative interventions can delay the growth of the so-called minimal residual disease. On the other hand, RCTs will be needed to rigorously answer the questions raised by preclinical and retrospective studies. While a single and short-term perioperative intervention (i.e., regional or general volatile anesthesia) is unlikely to reduce cancer recurrence or progression after oncologic surgery, it remains unknown whether the use combined therapies (i.e., beta blockers and NSAIDs) over a longer period of time that extends days before and after surgery can bring a much-needed benefit in the overall survival of patients with cancer.

Key Points

• 
  

  Cancer trials currently have one of the lowest enrollment rates among all clinical trials.

  
  
• 
  

  The future of perioperative cancer research will rely on improved collaboration, as well as conventional and novel study designs.

  
  
• 
  

  The identification of super-responders and nonresponders to cancer therapies will bring light to design novel strategies.

  
  
• 
  

  Basic and clinical researchers should take more responsibility to learn how to translate from bench to bedside and to design trials that address important clinical outcomes.