Health Condition 1: C00-D49- Neoplasms
Conditions
Interventions
None listed
Sponsors
AIRONC Healthcare Private Technologies Ltd
Eligibility
Inclusion criteria
Inclusion criteria: 1) Patient anonymized radiotherapy datasets of all disease sites: brain, head and neck, thorax (male and female), abdomen, pelvis (male and female) and extremity. 2) Diagnostic MRI, CT or PET scans used for radiotherapy contouring (both registered and non- registered datasets with planning CT). 3) Radiotherapy planning scans- initial and multiple scans take throughout the treatment (adaptive radiotherapy process). 4) Cone beam CT scans taken during radiotherapy treatment verification on linear accelerator machine. 5) Radiotherapy related data for all patients -- curative and palliative.
Exclusion criteria
Exclusion criteria: 1) Diagnostic scans of patients not treated with radiotherapy. 2) CT scans with non- standard spacing.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| Trained ML model with high accuracy scores (over 85%) on internal validation dataset with Dice similarity coefficient.Timepoint: 2 year | — |
Secondary
| Measure | Time frame |
|---|---|
| Accuracy of software by quantitative methods on test datasets: Dice similarity coefficient, Hausdorff distance, volume overlap error and relative volume difference.Timepoint: 1.5 years;Subjective score by oncologists: Five radiation oncologist will rate the auto-contours on a scale from 1-4 adopted from Lustberg et al and Liu et al. Timepoint: After Test validation;The dose received by manual and auto contours of organs at risk from the original treated plan will be recorded. Dose and volume parameters like mean dose, maximum dose and volume parameters.Timepoint: 2 years;Time for generating contours for each organ in every disease site for manual and auto in test datasets will be recorded in minutes.Timepoint: 2 years | — |
Countries
India
Contacts
Public ContactShirley Lewis
Kasturba Medical College Manipal
Outcome results
None listed