Reporting ‘Denominator’ data is essential for benchmarking and quality standards in ovarian cancer

Developing a platform to investigate the heterogeneity of outcomes for patients with ovarian cancer

BACKGROUND

The geographical variation in treatment patterns for patients with ovarian cancer is profound, long-standing and worrying. Although these variations were highlighted in a recent UK registry audit, granular data to provide explanations for these variations have been lacking.

METHODS

A consortium of six UK centres was generated to curate and submit data for all patients treated at their centre for a 2-year period. Descriptive statistics were combined with Cox regression and Kaplan-Meier analysis to confirm the findings from the national registry audit and identify possible drivers of the heterogeneity previously described.

RESULTS

Records for 1117 patients treated in six centres in 2018 and 2019 were collated. Although there were differences in the clinical characteristics of patients between centres, these were not enough to account for the significant variation in survival outcomes between centres (p<0.001). Treatment rates varied between centres with between 30% and 76% of patients receiving combination therapy but in Cox models 'treatment centre' remained a predictor of 1 year survival independent of patient, tumour factors and treatment choice.

CONCLUSION

Variations in outcome seen between UK centres are not related solely to casemix but rather to the approach and ethos of each centre towards advanced ovarian cancer treatment options. Although important, differences in treatment patterns do not completely explain the variations seen and further work is required to understand the drivers of difference seen.

British Gynaecological Cancer Society Recommendations for Evidence Based, Population Data Derived Quality Performance Indicators for Ovarian Cancer

Ovarian cancer survival in the UK lags behind comparable countries. Results from the ongoing National Ovarian Cancer Audit feasibility pilot (OCAFP) show that approximately 1 in 4 women with advanced ovarian cancer (Stage 2, 3, 4 and unstaged cancer) do not receive any anticancer treatment and only 51% in England receive international standard of care treatment, i.e., the combination of surgery and chemotherapy. The audit has also demonstrated wide variation in the percentage of women receiving anticancer treatment for advanced ovarian cancer, be it surgery or chemotherapy across the 19 geographical regions for organisation of cancer delivery (Cancer Alliances). Receipt of treatment also correlates with survival: 5 year Cancer survival varies from 28.6% to 49.6% across England. Here, we take a systems wide approach encompassing both diagnostic pathways and cancer treatment, derived from the whole cohort of women with ovarian cancer to set out recommendations and quality performance indicators (QPI). A multidisciplinary panel established by the British Gynaecological Cancer Society carefully identified QPI against criteria: metrics selected were those easily evaluable nationally using routinely available data and where there was a clear evidence base to support interventions. These QPI will be valuable to other taxpayer funded systems with national data collection mechanisms and are to our knowledge the only population level data derived standards in ovarian cancer. We also identify interventions for Best practice and Research recommendations.

Investigating the Impact of Ultra-Radical Surgery on Survival in Advanced Ovarian Cancer Using Population-Based Data in a Multicentre UK Study

We investigated URS and impact on survival in whole patient cohorts with AOC treated within gynaecological cancer centres that participated in the previously presented SOCQER 2 study. National cancer registry datasets were used to identify FIGO Stage 3,4 and unknown stage patients from 11 cancer centres that had previously participated in the SOCQER2 study. Patient outcomes’ association with surgical ethos were evaluated using logistic regression and Cox proportional hazards. Centres were classified into three groups based on their surgical complexity scores (SCS); those practicing mainly low complexity, (5/11 centres with >70% low SCS procedures, 759 patients), mainly intermediate (3/11, 35−50% low SCS, 356 patients), or mainly high complexity surgery (3/11, >35% high SCS, 356 patients). Surgery rates were 43.2% vs. 58.4% vs. 60.9%. across mainly low, intermediate and high SCS centres, respectively, p < 0.001. Combined surgery and chemotherapy rates were 39.2% vs. 51.8% vs. 38.3% p < 0.000 across mainly low, intermediate and high complexity groups, respectively. Median survival was 23.1 (95% CI 19.0 to 27.2) vs. 22.0 (95% CI 17.6 to 26.3) vs. 17.9 months (95% CI 15.7 to 20.1), p = 0.043 in mainly high SCS, intermediate, and low SCS centres, respectively. In an age and deprivation adjusted model, compared to patients in the high SCS centres, patients in the low SCS group had an HR of 1.21 (95% CI 1.03 to 1.40) for death. Mainly high/intermediate SCS centres have significantly higher surgery rates and better survival at a population level. Centres that practice mainly low complexity surgery should change practice. This study provides support for the utilization of URS for patients with advanced OC.

Factors Predicting Surgical Effort Using Explainable Artificial Intelligence in Advanced Stage Epithelial Ovarian Cancer

(1) Background: Surgical cytoreduction for epithelial ovarian cancer (EOC) is a complex procedure. Encompassed within the performance skills to achieve surgical precision, intra-operative surgical decision-making remains a core feature. The use of eXplainable Artificial Intelligence (XAI) could potentially interpret the influence of human factors on the surgical effort for the cytoreductive outcome in question; (2) Methods: The retrospective cohort study evaluated 560 consecutive EOC patients who underwent cytoreductive surgery between January 2014 and December 2019 in a single public institution. The eXtreme Gradient Boosting (XGBoost) and Deep Neural Network (DNN) algorithms were employed to develop the predictive model, including patient- and operation-specific features, and novel features reflecting human factors in surgical heuristics. The precision, recall, F1 score, and area under curve (AUC) were compared between both training algorithms. The SHapley Additive exPlanations (SHAP) framework was used to provide global and local explainability for the predictive model; (3) Results: A surgical complexity score (SCS) cut-off value of five was calculated using a Receiver Operator Characteristic (ROC) curve, above which the probability of incomplete cytoreduction was more likely (area under the curve [AUC] = 0.644; 95% confidence interval [CI] = 0.598−0.69; sensitivity and specificity 34.1%, 86.5%, respectively; p = 0.000). The XGBoost outperformed the DNN assessment for the prediction of the above threshold surgical effort outcome (AUC = 0.77; 95% [CI] 0.69−0.85; p < 0.05 vs. AUC 0.739; 95% [CI] 0.655−0.823; p < 0.95). We identified “turning points” that demonstrated a clear preference towards above the given cut-off level of surgical effort; in consultant surgeons with <12 years of experience, age <53 years old, who, when attempting primary cytoreductive surgery, recorded the presence of ascites, an Intraoperative Mapping of Ovarian Cancer score >4, and a Peritoneal Carcinomatosis Index >7, in a surgical environment with the optimization of infrastructural support. (4) Conclusions: Using XAI, we explain how intra-operative decisions may consider human factors during EOC cytoreduction alongside factual knowledge, to maximize the magnitude of the selected trade-off in effort. XAI techniques are critical for a better understanding of Artificial Intelligence frameworks, and to enhance their incorporation in medical applications.

The use of biomarkers to stratify surgical care in women with ovarian cancer: Scientific Impact Paper No. 69 March 2022: Scientific Impact Paper No. 69 May 2022

Biomarkers may offer unforeseen insights into clinical diagnosis, as well as the likely course and outcome of a condition. In this paper, the focus is on the use of biological molecules found in body fluids or tissues for diagnosis and prediction of outcome in ovarian cancer patients. In cancer care, biomarkers are being used to develop personalised treatment plans for patients based on the unique characteristics of their tumour. This tailoring of care can be used to pursue specific targets identified by biomarkers, or treat the patient according to specific tumour characteristics. Surgery is one of the core treatments for ovarian cancer, whether it is offered in primary surgery or following chemotherapy in delayed surgery. Biomarkers already exist to guide the treatment of tumours with chemotherapy, but very little research has determined the value of biomarkers in tailoring surgical care for ovarian cancer. Such research is required to identify new biomarkers and assess their effectiveness in a clinical setting as well as to help identify specific tumour types to guide surgery. Biomarkers could help to determine the success of removing the disease surgically, or help to identify tumour deposits that persist after chemotherapy. All of these aspects would improve current practice. This Scientific Impact Paper highlights research that may pave the way towards bespoke surgery according to the biological characteristics of a tumour and aid gynaecological oncologists to provide surgical treatment according to individual need, rather than a blanket approach for all.

Quality of life from cytoreductive surgery in advanced Ovarian cancer: investigating association with disease burden and surgical complexity in the international, prospective, SOCQER2 cohort study

OBJECTIVE

To investigate quality of life (QoL) and association with surgical complexity and disease burden after surgical resection for advanced ovarian cancer in centres with variation in surgical approach DESIGN: Prospective multicentre observational study SETTING: United Kingdom, Kolkata, India, and Melbourne, Australia gynaecological cancer surgery centres.

PARTICIPANTS

Patients undergoing surgical resection for late stage ovarian cancer. Exposure Low, intermediate or high Surgical Complexity Score (SCS) surgery MAIN OUTCOMES AND MEASURES: Primary: EORTC-QLQ-C30 Global score change. Secondary: EORTC OV28, progression free survival.

RESULTS

Patients' pre-operative disease burden and SCS varied between centres, confirming differences in surgical ethos. QoL response rates were 90% up to 18 months. Mean change from the pre-surgical baseline in the EORTC QLQ-C30 was 3.4 (SD 1.8, n=88) in the low, 4.0 (SD 2.1, n=55) in the intermediate and 4.3 (SD 2.1, n=52) in the high SCS group after 6 weeks (p=0.048) and 4.3 (SD 2.1, n=51), 5.1 (SD 2.2, n=41) and 5.1 (SD 2.2, n=35) respectively after 12 months (p=0.133). In a repeated measures model, there were no clinically or statistically meaningful differences in EORTC QLQ-C30 global scores between the three SCS groups, p= 0.840 but there was a small statistically significant improvement in all groups over time (p<0.001). The high SCS group experienced small to moderate decreases in physical (p=0.004), role (p=0.016) and emotional (p=0.001) function at 6 weeks post-surgery which resolved by 6-12 months.

CONCLUSIONS AND RELEVANCE

Global QoL of patients undergoing low, intermediate, and high SCS surgery improved at 12 months post operation and was no worse in patients undergoing extensive surgery.

Serum Albumin as a Predictor of Survival after Interval Debulking Surgery for Advanced Ovarian Cancer (AOC): A Retrospective Study

OBJECTIVE

To investigate the impact of serum albumin (at diagnosis and pre-operatively) on survival in patients undergoing cytoreductive surgery for advanced ovarian cancer(AOC) and whether improvement in albumin achieved following neoadjuvant chemotherapy (NACT) affects overall survival (OS).

METHODS

Outcomes of 441 patients who underwent cytoreduction for AOC were reviewed. Albumin was recorded at diagnosis and pre-operatively. Further analysis was performed if patients were hypoalbuminaemic at diagnosis.Analysis was stratified according to whether the patientreceived primary debulking surgery (PDS) or interval debulking surgery (IDS) and if their albumin was corrected.

RESULTS

308 patients had a serum albumin level at diagnosis and 400 patients had a pre-operative albumin available for analysis. For patients with an albumin at diagnosis ≤ 35g/L and ≥36 g/L, median OS was 31.5 (95% CI 23.5-39.5) and 50.4 (95% CI 38.9-61.9) months respectively (P = 0.003). Followingmultivariate analysis (MVA), albumin at diagnosis remained statistically significant as an independent marker for survival, even after adjusting for cytoreductive outcome, stage and grade(p = 0.04, Hazard ratio 1.38, 95% CI 1.01-1.89).Hypoalbuminaemic patients at diagnosis achieved complete cytoreduction in 53% of cases.For PDS patients, median OS was 19.7 months (95% CI 11.5-27.9). For IDS patients, median OS was 27.9 months (n = 1).IDS patients with a corrected albumin had a median OS of 42.9 months (95% CI 31.5-54.3) (p > 0.05).

CONCLUSION

Hypoalbuminaemia at diagnosis is a poor prognostic factor in AOC. Normalization of serum albumin after NACT is a potential predictor of survival.

Role of front-line bevacizumab in advanced ovarian cancer: the OSCAR study

OBJECTIVE

Two randomized phase III trials demonstrated the efficacy and safety of combining bevacizumab with front-line carboplatin/paclitaxel for advanced ovarian cancer. The OSCAR (NCT01863693) study assessed the impact of front-line bevacizumab-containing therapy on safety and oncologic outcomes in patients with advanced ovarian cancer in the UK.

METHODS

Between May 2013 and April 2015, patients with high-risk stage IIIB-IV advanced ovarian cancer received bevacizumab (7.5 or 15 mg/kg every 3 weeks, typically for ≤12 months, per UK clinical practice) combined with front-line chemotherapy, with bevacizumab continued as maintenance therapy. Co-primary endpoints were progression-free survival and safety (NCI-CTCAE v4.0). Patients were evaluated per standard practice/physician's discretion.

RESULTS

A total of 299 patients received bevacizumab-containing therapy. The median age was 64 years (range 31-83); 80 patients (27%) were aged ≥70 years. Surgical interventions were primary debulking in 21%, interval debulking in 36%, and none in 43%. Most patients (93%) received bevacizumab 7.5 mg/kg with carboplatin/paclitaxel. Median duration of bevacizumab was 10.5 months(range <0.1-41.4); bevacizumab and chemotherapy were given in combination for a median of three cycles (range 1-10). Median progression-free survival was 15.4 (95% CI 14.5 to 16.9) months. Subgroup analyses according to prior surgery showed median progression-free survival of 20.8, 16.1, and 13.6 months in patients with primary debulking, interval debulking, and no surgery, respectively. Median progression-free survival was 16.1 vs 14.8 months in patients aged <70 versus ≥70 years, respectively. The 1-year overall survival rate was 94%. Grade 3/4 adverse events occurred in 54% of patients, the most common being hypertension (16%) and neutropenia (5%). Thirty-five patients (12%) discontinued bevacizumab for toxicity (most often for proteinuria (2%)).

CONCLUSIONS

Median progression-free survival in this study was similar to that in the high-risk subgroup of the ICON7 phase III trial. Median progression-free survival was shortest in patients who did not undergo surgery.

Does sites of recurrence impact survival in secondary cytoreduction surgery for recurrent epithelial ovarian cancer?

Outcomes of secondary cytoreduction surgery (SCS) were evaluated for morbidity, progression free survival (PFS) and overall survival (OS) and factors influencing results were explored. Retrospective analysis of all cases of SCS for epithelial ovarian cancer (EOC) was performed from October 2010 to December 2017. 62 patients were prospectively identified as candidates for SCS and 57 underwent SCS. 20(35%) patients required bowel resection/s, 24(42%) had nodal resections and 11(19%) had extensive upper abdominal surgery. 51(89%) achieved complete cytoreduction. After a median follow-up of 30 months (range 9-95 months), median PFS was 32 months (CI 17-76 months) and median OS has not reached. Seventeen patients have died and 32 have progressed. Three patients had Clavien-Dindo grade-3 and two had grade-4 morbidity. Patients who had multi-site recurrence had shorter median PFS (p = 0.04) and patients who required bowel resections had lower median OS (p = 0.009) compared to rest of the cohort.IMPACT STATEMENTWhat is already known on this subject? Retrospective studies have confirmed survival advantage for recurrence in epithelial ovarian cancer and recommend SCS for carefully selected patients. This finding is being evaluated in randomised control trials currently.What do the results of this study add? This study presents excellent results for survival outcomes after SCS and highlights importance of careful selection of patients with a goal to achieve complete cytoreduction. In addition, for the first time in literature, this study also explores various factors that may influence results and finds that there are no differences in survival outcomes whether these patients had early stage or advanced stage disease earlier. Patients who have multisite recurrence tend to have shorter PFS but no difference were noted for overall survival. Patients who have recurrence in bowels necessitating resection/s have a shorter median OS compared to rest of cohorts, however, still achieving a good survival time.What are the implications of these findings for clinical practice and/or further research? These findings will raise awareness for the clinicians and patients while discussing surgical outcomes and would set an achievable standard to improve cancer services. The pattern of recurrence and associated outcomes also point towards difference in biological nature of recurrent disease and could provide an opportunity for scientists to study the biological makeup of these recurrent tumours.

Maximal-Effort Cytoreductive Surgery for Ovarian Cancer Patients with a High Tumor Burden: Variations in Practice and Impact on Outcome

BACKGROUND

This study aimed to compare the outcomes of two distinct patient populations treated within two neighboring UK cancer centers (A and B) for advanced epithelial ovarian cancer (EOC).

METHODS

A retrospective analysis of all new stages 3 and 4 EOC patients treated between January 2013 and December 2014 was performed. The Mayo Clinic surgical complexity score (SCS) was applied. Cox regression analysis identified the impact of treatment methods on survival.

RESULTS

The study identified 249 patients (127 at center A and 122 in centre B) without significant differences in International Federation of Gynecology and Obstetrics (FIGO) stage (FIGO 4, 29.7% at centers A and B), Eastern Cooperative Oncology Group (ECOG) performance status (ECOG < 2, 89.9% at centers A and B), or histology (serous type in 84.1% at centers A and B). The patients at center A were more likely to undergo surgery (87% vs 59.8%; p < 0.001). The types of chemotherapy and the patients receiving palliative treatment alone were equivalent between the two centers (3.6%). The median SCS was significantly higher at center A (9 vs 2; p < 0.001) with greater tumor burden (9 vs 6 abdominal fields involved; p < 0.001), longer median operation times (285 vs 155 min; p < 0.001), and longer hospital stays (9 vs 6 days; p < 0.001), but surgical morbidity and mortality were equivalent. The independent predictors of reduced overall survival (OS) were non-serous histology (hazard ratio [HR], 1.6; 95% confidence interval [CI] 1.04-2.61), ECOG higher than 2 (HR, 1.9; 95% CI 1.15-3.13), and palliation alone (HR, 3.43; 95% CI 1.51-7.81). Cytoreduction, of any timing, had an independent protective impact on OS compared with chemotherapy alone (HR, 0.31 for interval surgery and 0.39 for primary surgery), even after adjustment for other prognostic factors.

CONCLUSIONS

Incorporating surgery into the initial EOC management, even for those patients with a greater tumor burden and more disseminated disease, may require more complex procedures and more resources in terms of theater time and hospital stay, but seems to be associated with a significant prolongation of the patients overall survival compared with chemotherapy alone.

Socioeconomic differences impact overall survival in advanced ovarian cancer (AOC) prior to achievement of standard therapy

PURPOSE

Survival difference between socioeconomic groups with ovarian cancer has persisted in the United Kingdom despite efforts to reduce disparities in care. Our aim was to delineate critical episodes in the patient journey, where deprivation has most impact on survival.

METHODS

A retrospective review of 834 patients with advanced ovarian cancer (AOC) between 16/8/07-16/2/17 at a large cancer centre serving one of the most deprived areas of the UK. Using the Index of Multiple Deprivation (IMD), patients were categorised into five groups.

RESULTS

Surgery was more common in less deprived patients (p < 0.00001). Across IMD groups, there were no differences in complete (R0) cytoreduction rate (r = 0.18, p > 0.05), age, or comorbidity. The R0/total cohort rate increased with increasing IMD group (p < 0.0001). Patients refusing any intervention belonged exclusively to the three most deprived groups; 5/7 patients who refused surgery belonged to the most deprived IMD group. Overall survival in the total patient group was less in IMD group 1-2 compared to 9-10 (p = 0.002). On multivariate analysis, IMD group was not an independent predictor of survival (p > 0.05).

CONCLUSIONS

Socioeconomic differences in survival manifest in patients not receiving surgical treatment for AOC and are not purely explained by comorbidity, age, stage, or histological factors.

Enhancing ovarian cancer care: a systematic review of guideline adherence and clinical variation

Background

Clinical variation in ovarian cancer care has been reported internationally. Using Wennberg’s classification of clinical variation as effective care we can conceptualise variation through deviation from clinical guidelines. The aim of this review was to address knowledge gaps in the effectiveness of attempts to reduce unwarranted clinical variation through addressing the following questions: What is the evidence of guideline adherence in ovarian cancer and its deviation?; what are the key factors associated with variation in guideline adherence in ovarian cancer care?; and what quality improvement approaches have been used and what is the evidence of their effectiveness in enhancing guideline adherence in ovarian cancer care?.

Methods

Keywords and synonyms for the major concepts of ovarian cancer, guideline adherence and safety were developed and combined to form the search strategy. Systematic searches of four electronic databases were undertaken of publications from January 2007 to November 2018. Retrieved articles were assessed against the eligibility criteria to determine those for inclusion.

Results

Thirty-two papers were included in the review with three broad groupings identified: adherence to and deviation from guidelines (either local, national or international guidelines); factors impacting guidelines adherence; and quality improvement approaches.

Conclusions

Unwarranted clinical variation may be used as a marker for the effectiveness of a health system, based on the outcome of this systematic review. This review found that the implementation of quality indicators through a formal quality improvement program lead to improvements in guideline adherent care. Further research on outcomes of implementing quality improvement programs in ovarian cancer care will improve the ability to implement centralised care and further identify factors that to improve outcomes in ovarian cancer care.

Electronic supplementary material

The online version of this article (10.1186/s12889-019-6633-4) contains supplementary material, which is available to authorized users.

The NICE classification for 'Ultra-radical (extensive) surgery for advanced ovarian cancer' guidance does not meaningfully predict postoperative complications: a cohort study

OBJECTIVE

To determine which descriptors of cytoreductive surgical extent in advanced ovarian cancer (AOC) best predict postoperative morbidity.

DESIGN

Retrospective notes review.

SETTING

A gynaecological cancer centre in the UK.

POPULATION

Six hundred and eight women operated on for AOC over a period of 114 months at a tertiary cancer centre, between 16 August 2007 and 16 February 2017.

METHODS

Outcome data were analysed by six approaches to classify the extent of surgery: standard/ultra-radical surgery; standard/radical/supra-radical surgery; presence/absence of gastrointestinal resections; low/intermediate/high surgical complexity score (SCS); presence of bowel anastomoses and/or diaphragmatic surgery; and the presence/absence of multiple bowel resections.

MAIN OUTCOME MEASURES

Major (grades 3-5) postoperative morbidity and mortality.

RESULTS

Forty-three (7.1%) patients experienced major complications. Grade-5 complications occurred in six patients (1.0%). Patients who underwent multiple bowel resections had a relative risk (RR) of 7.73 (95% confidence interval, 95% CI 3.92-15.26), patients with a high SCS had an RR of 6.12 (95% CI 3.25-11.52), patients with diaphragmatic surgery and gastrointestinal anastomosis had an RR of 5.57 (95% CI 2.65-11.72), patients with 'any gastrointestinal resection' had an RR of 4.69 (95% CI 2.66-8.24), patients with ultra-radical surgery had an RR of 4.65 (95% CI 2.26-8.79), and patients with supra-radical surgery had an RR of 4.20 (95% CI 2.35-7.51) of grades 3-5 morbidity, compared with patients undergoing standard surgery as defined by the National Institute for Health and Care Excellence (NICE) in the UK. No significant difference was seen in the rate of major morbidity between standard (6/59, 10.2%) and ultra-radical (9/81, 11.1%) surgery within the cohort who had intermediate complex surgery (P > 0.05).

CONCLUSIONS

The numbers of procedures performed significantly correlate with major morbidity. The number of procedures performed better predicted major postoperative morbidity than the performance of certain 'high risk' procedures. We recommend using SCS to define a higher risk operation. NICE should re-evaluate the use of the term 'ultra-radical' surgery.

TWEETABLE ABSTRACT

Multiple bowel resection is the best predictor of morbidity and is more predictive than 'ultra-radical surgery'.