Copenhagen, Denmark
Onsite/Online

ESTRO 2022

Session Item

Sunday
May 08
09:00 - 10:00
Mini-Oral Theatre 1
09: Personalised radiation therapy
Brita Singers Sørensen, Denmark;
Rita Simoes, United Kingdom
2160
Mini-Oral
Interdisciplinary
Saliva microbiota and inflammation markers predict acute toxicity after RT for head-and-neck cancer
Ester Orlandi, Italy
MO-0381

Abstract

Saliva microbiota and inflammation markers predict acute toxicity after RT for head-and-neck cancer
Authors:

Nicola Alessandro Iacovelli1, Tiziana Rancati2, Rossana Ingargiola1, Salvatore Alfieri3, Loris De Cecco4, Fabio Badenchini2, Anna Cavallo5, Alessandro Cicchetti2, Nadia Zaffaroni6, Valentina Doldi6, Elisa Mancinelli4, Mara Serena Serafini4, Andrea Devecchi4, Riccardo Valdagni7, Ester Orlandi1

1Fondazione IRCCS Istituto Nazionale Tumori , Division of Radiation Oncology 2, Milan, Italy; 2Fondazione IRCCS Istituto Nazionale Tumori , Prostate Cancer Program, Milan, Italy; 3Fondazione IRCCS Istituto Nazionale Tumori , Division of Medical Oncology 3, Milan, Italy; 4Fondazione IRCCS Istituto Nazionale Tumori , Department of Applied Research and Technology Development, Milan, Italy; 5Fondazione IRCCS Istituto Nazionale Tumori , Division of Medical Physics, Milan, Italy; 6Fondazione IRCCS Istituto Nazionale Tumori , Division of Molecular Pharmacology, Milan, Italy; 7Fondazione IRCCS Istituto Nazionale Tumori , Division of Radiation Oncology 1, Milan, Italy

Show Affiliations
Purpose or Objective

We hypothesised that the saliva microbiota (MB) and cytokine levels before radiotherapy (RT) differ between patients with/wout acute toxicity (tox) after RT for and head&neck (HNC) cancer

Material and Methods

We enrolled 114 consecutive HNC pts treated with conventional (54-70Gy @2Gy/fr) or moderately hypofractionated (46.6-69.9Gy @2.1-2.2Gy/fr) VMAT+IGRT.

 

A detailed evaluation was done pre-, during & at RT end, including saliva MB measures (16S sequencing and pooling in Operational Taxonomic Units -OTUs- with Uclust software) and saliva assessment of cytokines  (TNα, IL1b according to previous literature, Bossi2016).

 

Tox was scored weakly using CTCAE; we chose a longitudinal definition of tox, taking both severity & duration into account. Average grade>1.5 for oral mucositis during RT (aOM) was the endpoint for this analysis

 

We used logistic regression (LR) to derive inflammation signatures (based on cytokine levels at baseline) and unsupervised clustering (fuzzy c-means) to partition pts into MB clusters based on the relative abundance of OTUs before RT start.

 

Information on inflammation & MB clustering was introduced in a sigmoid-shaped dosimetric a Normal Tissue Complication Probability (NTCP) model to test their added value.

Results

Toxicity was scored in 22/114 pts.

The baseline concentration of IL6  was significantly associated with acute tox: OR=1.8 (continuous log-scale, p=0.05). Fig1a presents results when stratifying at 33°-66° percentiles OR=1.8 for each step (p=0.04). We defined a favourable IL6 profile if IL6 in the lowest 10° percentile (logIL6 (ng/ml) <0.7), 15.4 vs 36.8% aOM in favourable vs unfavourable IL6.

MB clustered in 2 groups at the Genus level, with 9 genera included in the centroid signature (Fig1b). With Haemophilus, Neisseria, Prevotella and Streptococcus mostly driving the pts grouping. Pts in cluster B had a significantly higher probability of aOM (unfavourable MB) compared to pts in cluster A (favourable MB): tox rates were 22.7 vs 14.6%, OR=1.7 (p=0.05).

MB clustering was confirmed in the validation cohort: tox rates 19 vs 32% in unfavourable vs favourable MB (without any change in centroids for clustering).

To join information from MB and inflammation marker, we classified pts at low-risk (LR) of tox if they had “favourable MB AND IL6 profile”, at intermediate-risk (IR) if “favourable MB OR IL6 profile”, at high-risk (HR) if “unfavourable MB AND IL6 profile”.  Observed toxicity rates in LR/IR/HR were 12.5/16.7/41.2% (p=0.04).

We obtained different tolerance doses for different risk classes when including “biological” stratification into a NTCP model (Fig.2).




Conclusion

We determined 3 risk classes for RT-induced acute side effects based on the combination of MB information and cytokine profile. The biologically personalised risk prediction improves discrimination and allows to the design of possible interventional trials to reduce tox by modifying MB/inflammation levels before RT starts.


This research was funded under the Call for the Promotion of Institutional Research INT year 2016, 5 X 1000 Italian Ministry of Health