Improved surgical treatment of oesophageal cancer

Abstract

Oesophageal cancer is the 7thmost common cancer globally and the 5-year survival is poor (below 20%). Curative treatment usually involves surgical resection of the tumour (oesophagectomy), with or without neoadjuvant chemo(radio)therapy. The aim of the thesis was to identify surgery-related factors of importance for improved long-term survival in oesophageal cancer.

Study I was a nationwide Swedish cohort study of patients who underwent oesophagectomy for oesophageal cancer between 1987 and 2010, with follow-up until 2016. The study included 1,384 patients who had undergone surgery by any of 36 surgeons. Risk adjusted cumulative sum analysis was used to create proficiency gain curves for “lower volume surgeons” (<4 cases per year) and “higher volume surgeons” (≥4 cases per year), as well as “younger surgeons” (<45 years) and “older surgeons” (≥45 years) regarding all-cause 1 to 5-year mortality (main outcome). The results were adjusted for confounders. “Higher volume surgeons” reached proficiency at 14 cases compared to 31 cases for “lower volume surgeons”. “Younger surgeons” reached proficiency at 13 cases compared to 48 cases for “older surgeons”.

Study II was a systematic review and meta-analysis comparing long-term survival after minimally invasive oesophagectomy (MIO) with open oesophagectomy (OO) for oesophageal cancer in studies published up until 2018. Based on 55 relevant studies and 14,592 patients (7,358 MIO and 7,234 OO), random effects meta-analysis was used to produce hazard ratios (HR) with 95% confidence intervals (CI) for all-cause 5-year mortality (main outcome) with adjustment for confounders. MIO was associated with 18% lower risk of all-cause 5-year mortality compared to OO (HR 0.82, 95% CI 0.76-0.88).

Study III was a population-based cohort study including almost all patients operated for oesophageal cancer in Sweden from 2011 until 2015 and in Finland from 2010 until 2016, with follow-up throughout 2019. Multivariable Cox regression was used to produce HRs with 95% CIs comparing MIO (n=459) with OO (n=771) for the main outcome all-cause 5-year mortality. The results were adjusted for confounders. MIO was associated with 18% lower risk of all-cause 5-year mortality compared to OO (HR 0.82, 95% CI 0.67-1.00 [P=0.048]).

Study IV was a population-based cohort study including almost all patients who underwent surgery for oesophageal cancer from 2000 until 2015 in Sweden and from 2000 until 2016 in Finland, with follow-up throughout 2019. The 2,306 included patients were divided into deciles (10 about equal size group) by the level of lymphadenectomy during oesophagectomy. Multivariable Cox regression was used to produce HRs with 95% CIs for the main outcome all-cause 5-year mortality with adjustment for confounders. Compared to the 1st decile (0-3 nodes) the lowest risk for all-cause 5-year mortality was found in decile 8 (25-30 nodes). Upon stratification, this survival benefit was especially apparent for T3/T4 tumours and for patients who did not receive neoadjuvant therapy.

In conclusion, this thesis indicates that intense training in oesophagectomy of younger surgeons, use of minimally invasive oesophagectomy and moderate extent of lymphadenectomy improve long-term survival in patients who undergo surgery for oesophageal cancer.