Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Investigation of the prognostic value of CT and PET-based radiomic image features in oropharyngeal squamous cell carcinoma

Said, Mohammed Mosad LU (2016) MSFT01 20161
Medical Radiation Physics, Lund
Medical Physics Programme
Abstract
Background
Medical data in the form of radiographic routine scans is steadily accumulating. The analysis of such data through automated quantitative methods is believed to produce new information which would allow for more personalization of therapy. The present thesis investigated the use of such methods in head and neck cancer.

Material and methods
Pretreatment positron emission tomography (PET) and computed tomography (CT) scans from 74 patients present with oropharyngeal squamous cell carcinoma were analyzed quantitatively and a total of 92 image-based features were calculated. These features attempt to describe the shape and size of the tumor, as well as the heterogeneity within. The prognostic value of these features, as well as... (More)
Background
Medical data in the form of radiographic routine scans is steadily accumulating. The analysis of such data through automated quantitative methods is believed to produce new information which would allow for more personalization of therapy. The present thesis investigated the use of such methods in head and neck cancer.

Material and methods
Pretreatment positron emission tomography (PET) and computed tomography (CT) scans from 74 patients present with oropharyngeal squamous cell carcinoma were analyzed quantitatively and a total of 92 image-based features were calculated. These features attempt to describe the shape and size of the tumor, as well as the heterogeneity within. The prognostic value of these features, as well as common clinical variables, was investigated for tumor recurrence and disease-specific mortality, respectively. Additionally, prediction of treatment failure was attempted using an artificial neural network. All patients received intensity-modulated radiation therapy and there were thus treatment plans for each patient in addition to the PET/CT scans. The non-uniformity of the dose distribution was studied using custom features based on the gray-level size zone matrix. These custom features measured the number of disconnected regions receiving either too low or too high of radiation dose, and differences in the sizes of such regions.

Results
One PET- and two CT-based features were found to significantly differ between responders and non-responders. The PET-based feature was the correlation (p = 0.0011), which is a texture feature derived from the gray-level co-occurrence matrix. It described the irregularity in radiotracer uptake on a voxel-to-voxel basis and results suggest that non-responders have more irregular patterns of uptake. The CT-based features were the variance (p = 0.0012) and skewness (p = 0.0027), where the former was found to be significantly larger among responders and the skewness more negative. However, image-based features performed quite poorly in treatment failure prediction, as compared to clinical variables, which had an area under the receiver operating characteristic curve (AUC-ROC) of 0.87 (95% confidence interval, 0.73—0.96) for primary tumor recurrence and 0.73 (95% confidence interval, 0.52—0.87) for disease-specific mortality. Three image-based features did, however, contribute significantly when included to the model utilizing clinical variables, which suggests that they may contain additional information that is likely to be of value. Of the five custom features calculated on the dose distribution, the one emphasizing differences in the number of disconnected regions was observed to be significantly higher among non-responders. No statistical differences were found in the sizes of low-dose and high-dose regions, respectively, between the two groups.

Conclusion
Quantitative analysis of routine scans may provide additional information regarding tumor phenotype, which is likely to be of value when used in conjunction with clinical variables. Additionally, texture analysis of the dose distribution reveals differences between treatment plans that are not captured by dose-volume histogram metrics. These methods are, however, relatively new in use on medical data and there are certain details that require further investigation. (Less)
Popular Abstract (Swedish)
En av de stora utmaningarna inom onkologin är den stora heterogeniteten mellan olika tumörer. Ingen tumör är den andra lik och detta skapar ett behov av individualiserad vård där behandlingen skräddarsys efter patienten. I regel grundas behandlingsalternativen på populationsbaserad forskning och den givna behandlingen är därmed optimal för en så kallad ”medelpatient”, men en sådan existerar inte och i verkligheten är det vissa som överbehandlas medan andra underbehandlas. Stora framsteg har gjorts i denna individualisering genom rutinmässiga vävnadsprover och identifieringen av ett stort antal biomarkörer. Studier har dock visat på stora variationer i genuttryck inom en och samma tumör och nackdelen med vävnadsprover är att de återger... (More)
En av de stora utmaningarna inom onkologin är den stora heterogeniteten mellan olika tumörer. Ingen tumör är den andra lik och detta skapar ett behov av individualiserad vård där behandlingen skräddarsys efter patienten. I regel grundas behandlingsalternativen på populationsbaserad forskning och den givna behandlingen är därmed optimal för en så kallad ”medelpatient”, men en sådan existerar inte och i verkligheten är det vissa som överbehandlas medan andra underbehandlas. Stora framsteg har gjorts i denna individualisering genom rutinmässiga vävnadsprover och identifieringen av ett stort antal biomarkörer. Studier har dock visat på stora variationer i genuttryck inom en och samma tumör och nackdelen med vävnadsprover är att de återger endast en liten del av tumören och därmed misslyckas med att fånga denna heterogenitet.

Ytterligare ett framsteg är den rutinmässiga bildtagningen med datortomografi (CT), som ger en god helhetsbild av tumörstorleken samt spridningen i kroppen. På senare tid har även molekylär bildtagning blivit rutin och information gällande tumörfunktion har blivit tillgänglig. Utifrån dessa bilder görs en någorlunda kvalitativ bedömning som bidrar enormt till behandlingsstrategin. En kvantitativ analys av tumören, som sedd på bilder tagna innan behandlingsstart, kan dock mynna i ny information som har visats vara användbar för att prediktera behandlingsrespons. Analysen är automatiserad och syftar på att undersöka storleken så väl som formen på tumören, men också den inre heterogeniteten, vilket beroende på bildmodalitet kan spegla olika saker.

Det finns idag en enorm mängd data i form av medicinska bilder som bara ökar med tiden. Radiomics är ett relativt nytt ämnesområde där dessa bilder analyseras genom kvantitativa metoder i syftet att utvinna värdefull information, som kan användas i individualiseringen av framtida behandlingar. Den stora mängden data vidare möjliggör användandet av avancerade maskininlärningsmetoder för bättre risk-stratifiering.
I detta examensarbete undersöktes det prognostiska värdet av ett relativt stort antal kvantitativa parametrar beräknade på både CT och positron emission tomography (PET). Av de totalt 92 beräknade parametrarna var det en PET-baserad och två CT-baserade parametrar som lyckades differentiera mellan patienter med positiv och negativ behandlingsrespons. Utöver det användes ett artificiellt neuronnät för att prediktera behandlingsrespons utifrån både bildbaserade parametrar och kliniska variabler. Med denna information, som är tillgänglig redan innan behandlingsstart, kunde responsen predikteras med ganska god säkerhet, vilket kan vara värdefullt ur klinisk synpunkt eftersom alternativa behandlingsstrategier kan utforskas i ett tidigt stadium.

Inom strålbehandling finns dessutom information om dosfördelningen, som bör uppfylla angivna doshomogenitetskrav. Uniformiteten av dosfördelningen analyserades med hjälp av fem speciellt framtagna parametrar, som tog hänsyn till det spatiala förhållandet mellan både lågdos- och högdosområden. Resultaten visade att antalet osammanhängande lågdos- och högdosområden kan vara av betydelse för behandlingsrespons. Detta är dock en hittills obeprövad metod för analys av dosfördelningar och det finns utrymme för förbättringar. (Less)
Please use this url to cite or link to this publication:
author
Said, Mohammed Mosad LU
supervisor
organization
course
MSFT01 20161
year
type
H2 - Master's Degree (Two Years)
subject
language
English
id
8897993
date added to LUP
2017-01-08 17:09:59
date last changed
2017-01-09 16:29:38
@misc{8897993,
  abstract     = {{Background
Medical data in the form of radiographic routine scans is steadily accumulating. The analysis of such data through automated quantitative methods is believed to produce new information which would allow for more personalization of therapy. The present thesis investigated the use of such methods in head and neck cancer.

Material and methods
Pretreatment positron emission tomography (PET) and computed tomography (CT) scans from 74 patients present with oropharyngeal squamous cell carcinoma were analyzed quantitatively and a total of 92 image-based features were calculated. These features attempt to describe the shape and size of the tumor, as well as the heterogeneity within. The prognostic value of these features, as well as common clinical variables, was investigated for tumor recurrence and disease-specific mortality, respectively. Additionally, prediction of treatment failure was attempted using an artificial neural network. All patients received intensity-modulated radiation therapy and there were thus treatment plans for each patient in addition to the PET/CT scans. The non-uniformity of the dose distribution was studied using custom features based on the gray-level size zone matrix. These custom features measured the number of disconnected regions receiving either too low or too high of radiation dose, and differences in the sizes of such regions.

Results
One PET- and two CT-based features were found to significantly differ between responders and non-responders. The PET-based feature was the correlation (p = 0.0011), which is a texture feature derived from the gray-level co-occurrence matrix. It described the irregularity in radiotracer uptake on a voxel-to-voxel basis and results suggest that non-responders have more irregular patterns of uptake. The CT-based features were the variance (p = 0.0012) and skewness (p = 0.0027), where the former was found to be significantly larger among responders and the skewness more negative. However, image-based features performed quite poorly in treatment failure prediction, as compared to clinical variables, which had an area under the receiver operating characteristic curve (AUC-ROC) of 0.87 (95% confidence interval, 0.73—0.96) for primary tumor recurrence and 0.73 (95% confidence interval, 0.52—0.87) for disease-specific mortality. Three image-based features did, however, contribute significantly when included to the model utilizing clinical variables, which suggests that they may contain additional information that is likely to be of value. Of the five custom features calculated on the dose distribution, the one emphasizing differences in the number of disconnected regions was observed to be significantly higher among non-responders. No statistical differences were found in the sizes of low-dose and high-dose regions, respectively, between the two groups.

Conclusion
Quantitative analysis of routine scans may provide additional information regarding tumor phenotype, which is likely to be of value when used in conjunction with clinical variables. Additionally, texture analysis of the dose distribution reveals differences between treatment plans that are not captured by dose-volume histogram metrics. These methods are, however, relatively new in use on medical data and there are certain details that require further investigation.}},
  author       = {{Said, Mohammed Mosad}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Investigation of the prognostic value of CT and PET-based radiomic image features in oropharyngeal squamous cell carcinoma}},
  year         = {{2016}},
}