• Orthopaedic oncology

Background
Soft tissue sarcomas (STS) are rare malignancies, accounting for 1% of adult cancers. There are
over 70 histological subtypes with a 5-year overall survival (OS) of 59–78% and a 5-year local
recurrence (LR) rate from 12% to 28%, depending on the tumor and patient characteristics. The
most typical location is the extremities, and treatment typically involves surgery and Radiotherapy
(RT). Wound complications occur in up to 56% of patients. The overall aim of this thesis was to
investigate outcomes for Danish patients with deep, high-grade STS of the extremities and trunk
wall, focusing on the impact of the Danish Cancer Patient Pathway (CPP), differences in RT
practices, and predictive survival modeling. Additionally, we present the rationale and study
design of a Randomized control trial (RCT) with the aim of exploring the potential of negative
pressure wound therapy (NPWT) to reduce surgical wound complications. However, the final
inclusion in the RCT has not been finalized.
Methods
This thesis is based on four cohorts: two retrospective national cohorts of patients with deep, high
malignant STS of the extremities and trunk wall diagnosed between 2000 and 2016. The third
cohort is a national retrospective cohort with deep, high-malignant STS of the extremities and
trunk wall diagnosed between 2000 and 2018, excluding patients diagnosed in 2009, when the
Danish CPP was implemented. The last cohort is a single-center prospective cohort, including
patients receiving surgery for deep high malignant STS at Rigshospitalet from 2020 forward.
Main Results
In Study I we found that the 5-year overall survival rate for the entire cohort (2000-2018) was 48%
(95% CI: 45–52). When we divided the cohort into pre-CPP (2000-2008), we found a 5-year survival
rate of 43% (95% CI: 38–49), and the post-CPP (2010-2018) was 52% (95% CI: 47–57) (p=0.05). The
pre-CPP cohort had a significantly higher rate of unplanned sarcoma operation at non-specialized
centers (‘whoops procedures’) than the post-CPP cohort (10% vs. 4%, p<0.001). The post-CPP
cohort also had a shorter diagnostic-to-treatment interval, with a median of 15 days (IQR: 9.25–
11
21) versus 18 days (IQR: 12–30) in the pre-CPP cohort (p<0.001). In study II (2000-2016), we found
that 83% of the patients treated at Aarhus University Hospital (AUH) received postoperative RT
compared to 64% at Rigshospitalet (RH) (p<0.001). There was no difference in the five-year
cumulative incidence of LR; AUH had 15% (95% CI: 10–20%) and RH had 14% (95% CI: 9–19%)
(p=0.6). In study III, we developed and internally validated a Logistic Regression model for fiveyear
survival with an AUC of 0.74 (95% CI: 0.66–0.82), a sensitivity of 0.65, and specificity of 0.72,
together with strong predictive performance (PPV: 0.65, NPV: 0.72) and a Brier score of 0.41. In
comparison, Random Forest (RF) had an AUC of 0.65 (95% CI: 0.56–0.74), sensitivity of 0.59,
specificity of 0.69, PPV of 0.60, NPV of 0.68, and Brier score of 0.38. Study IV is a protocol article
for an ongoing single-center RCT evaluating NPWT in wound healing after surgery for high-grade,
deep-seated STS. Patients are randomized 1:1 to closure with staples and either conventional
dressing or NPWT, with a one-year follow-up. The study, powered at 80% (α = 0.05, β = 0.20),
estimates an 80% reduction in wound complications based on retrospective data. The target
sample is 160 patients (80 per group) to account for dropouts. The study is ongoing, and the
results are still pending.
Conclusion
We demonstrated that implementing the CPP for sarcomas improved patient survival and reduced
waiting times from diagnosis to treatment. Additionally, the CPP contributed to a significant
reduction in Whoops procedures. Furthermore, we found that a difference in the two sarcoma
centers’ approaches to adjuvant RT visitation did not result in higher five-years LR rates,
suggesting that RT can be safely omitted in select patient groups operated on with wide margins
without compromising outcomes. Our predictive models demonstrated potential for personalized
survival estimates based on patient and tumor characteristics, with the Logistic regression model
demonstrating a slightly better overall performance on our data. Finally, our forthcoming RCT
results aim to enhance postoperative wound management. We hope to emphasize the importance
of individualized patient-specific treatment strategies in STS with our findings.