Artificial Intelligence-Enabled Skin Perforator Segmentation
- Conditions
- Microvascular Free Flap TransferFibula Flap
- Interventions
- Other: AI
- Registration Number
- NCT06634472
- Lead Sponsor
- The University of Hong Kong
- Brief Summary
Computer-assisted surgery (CAS) has revolutionised head and neck reconstruction with more efficient, accurate, and predictable surgery, as reported in our previous studies. Skin perforators are perforating vessels that travel through muscles and septa to supply the skin. The identification of skin perforators is crucial for a safe fibula osteocutaneous free flap (FFF) harvest with CAS. Different methods have been proposed in the past, each of which has its own limitations.
Traditionally, skin perforators are identified with a Doppler ultrasound. Berrone et al. measured the locations with a Doppler and imported the information back to guide virtual surgical planning (VSP). However, their study showed imprecise concordance between handheld Doppler measurements and the actual perforator locations; good correlation between the location of perforators and bone segments was identified in only four out of six cases investigated. To improve on the accuracy, computed tomography angiography (CTA) was used for skin perforator identification. Battaglia et al. manually marked the perforating vessel location at the subcutaneous level and reported good correlation. However, the manual segmentation of the perforator was at the subcutaneous level only. The course of the perforators, which would be more significant for the design of CAS FFF harvest was not shown.
To incorporate the course of skin perforators into FFF VSP, Ettinger et al. first described the technique of manual tracing from CTA in 2018 and validated its accuracy in 2022. The median absolute difference between the CTA and intraoperative measurements was 3 mm. However, reports quoted an average of 2 to 3 hours spent on tracing and modelling the course of the perforators depending on their number and anatomy; consequently this adds significant burden to healthcare professionals.
Recently, United imaging intelligence has developed an AI-based program that offers a potential solution for accurate and efficient localisation of skin perforators to be incorporated into the current VSP workflow. The proposed study aims to validate its performance in a prospective case series. This will be the first study to investigate the use of AI-enabled program for fibula skin paddle perforator identification.
- Detailed Description
Aims and Hypotheses to be Tested:
The aim of this single-arm clinical trial is to validate the performance of the AI-enabled skin perforator segmentation tool (AI tool) in computer-assisted FFF harvest.
To test the predictive accuracy (PA) of the AI tool in identifying the targeted skin perforators (primary endpoint)
* The primary outcome of the PA is calculated from the number of true/false positive (TP/FP), true/false negative (TN/FN) targeted perforators.
* The secondary outcomes include the sensitivity (SEN), specificity (SPE), positive, and negative predictive values (PPV and NPV respectively).
Our hypothesis is that the AI-enabled vessel segmentation tool will demonstrate a high level of accuracy in identifying the skin perforators during the CAS osteocutaneous FFF harvest.
Plan of Investigation:
(i) Study design
The current study is a single-arm single-centre clinical trial to explore the performance of an AI-enabled skin perforator segmentation tool in CAS FFF harvest. The study protocol shall conform to Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) and Practical Robust Implementation and Sustainability Model (PRISM).
Data collection procedure A case report form (CRF attached in the Appendix) has been designed to collect pre- and intra-operative research data.
Patient demographics will be collected, including age, gender, body height, weight, body mass index (BMI), diagnosis, medical history, and smoking status.
Preoperative measurements after incorporating the AI-segmented skin perforators to the VSP will also be obtained, including but not limited to:
* Total fibula length
* Number and lengths of fibula segments
* Number of immediate dental implants
* Number of perforators
* Type of perforators:
1. Septocutaneous vs musculocutaneous
2. Targeted (to be incorporated into the VSP and harvesting) vs accessory (other perforators)
* Relative location of the perforators, measured as the distance from the most inferior point of the lateral malleolus to:
1. Level of the perforator, as it crosses the posterior aspect of the fibula
2. The point where it bifurcates from the peroneal artery
* Time spent incorporating the skin perforator model to the original VSP workflow
Intraoperatively, the following information will be recorded in CRF:
* Deviation between the locations of the skin perforators identified by the AI tool and during the surgery at the point, where it crosses the posterior edge of the fibula
* Need for intraoperative modifications of the VSP due to the perforator identification discrepancy
* Operator satisfaction with the AI-segmentation tool
The primary endpoint is the predictive accuracy of the AI-enabled skin perforator segmentation tool. It will be calculated according to the definition by Šimundić et al. When the skin perforator is identified by both the AI-segmentation tool and during the surgery, it will be counted as a true positive (TP). When the perforator is identified by the AI-segmentation tool, but not found during the surgery it is counted as a false positive (FP). When the perforator is seen during the surgery, but not shown by the AI tool, it is a false negative (FN). Finally, a true negative (TN) perforator count will be derived from those subjects, who do not exhibit an FN perforator. The predictive accuracy (PA) is identified as the percentage of true perforators among all the perforators, and is calculated as (TP +TN)/(TP+FP+TN+FN)\*100%.
The secondary outcomes of sensitivity (SEN), specificity (SPE), positive and negative predictive values (PPV and NPV, respectively) of the AI segmentation tool will be calculated according to the definition given in Table 1. These outcomes will be calculated at the perforator trunk for all the perforators when the perforators cross the posterior edge of the fibula, and at the branches for the targeted perforators when they enter the subcutaneous layer of skin.
(ii) Methods
Prof. Pu and Prof. Su have been working on computer assisted jaw reconstruction, virtual surgical planning for several years. The workflow has already been set up for the previously funded HMRF projects. The efficiency, predictability, and accuracy of the protocol have been demonstrated in previously published articles.
For the currently proposed project, the investigators aim to incorporate the AI tool into the current VSP workflow and assess its performance. The complete process will be divided into three parts: 1) preoperative preparation; 2) surgery and intra-operative recording; 3) postoperative revaluation. In the preliminary study, the investigators have established the feasibility of the methodology via several pilot cases.
1. Preoperative preparations:
Preoperative computed tomography (CT) imaging Patients enrolled in the study will be arranged to undergo a CT scan on a General Electric revolution system (GE Healthcare, Chicago, IL) with slice thicknesses of 0.625 mm for both the head and neck, and lower limbs. Iodinated contrast medium (Iopamiro 370, Bracco, Israel) at a concentration of 370 mg/ml will be administered intravenously at a flow rate of 4 to 5 ml/sec, CT angiogram of the lower extremity will be taken at the arterial phase. The imaging data will be exported in the Digital Imaging and Communications in Medicine (DICOM) format.
AI-enabled segmentation of skin perforators from CT angiogram The DICOM data of the CTA will be imported to the AI workstation (United Imaging Intelligence (UII), Shanghai, China). The tibia and fibula bone, skin, blood vessels of the lower extremities including the anterior tibial artery (ATA), posterior tibial artery (PTA), and peroneal artery (PEA) will be segmented with pre-set threshold range to generate the stereolithography (.stl) models. Based on the fully CNN of V-net, the blood vessel branches from the PEA will be automatically segmented to build the .stl model. The small perforating vessels from the PEA traveling laterally crossing the posterior edge of the fibula bone to the skin surface will be marked as skin perforators. The .stl models of the fibula bone, skin, ATA, PTA, PEA, and skin perforators will be exported from the workstation for incorporation into the VSP workflow.
Virtual surgery planning
Preoperative imaging and building of models Preoperative CT scan will be imported to ProPlan CMF 2.1 software (Materialise, Leuven, Belgium). Volume segmentation will be performed to build the 3D virtual models of the jaws.
Incorporating the skin perforators The .stl models exported from the AI workstation will be imported to the ProPlan file. The fibula length and skin perforator location will be measured and recorded.
Virtual fibula flap harvest, segmentation, and dental implant insertion VSP will be performed using the methodology described in the previous publications. The extent of resection will be pre-determined, based on the type and extension of the pathology. The osteotomies for fibula harvest and segmentation will be planned to suit the defect after guided tumour resection. The location of the immediate dental implants will be marked in the fibula segments. Taking the location of the skin perforators into consideration, the osteotomies can be adjusted to include the take-off point of the skin perforators, while avoiding placing the perforator directly in planned osteotomies or implant insertion points.
Design and fabrication of surgical guides and/or plates After the VSP is confirmed, the surgical guides and/or plates will be designed using 3-matic 13.0 software (Materialise, Leuven, Belgium). The location of the perforators as they cross the posterior edge of the fibula will be marked with a small sphere on the fibula harvest guide. The guides will be printed with biocompatible and autoclavable resin - either MED610 (Stratasys Ltd, Eden Prairie, MN, USA) or NextDent SG (Vertex Dental, The Netherlands).
2. Surgical procedures and intraoperative recording All the surgical procedures will be conducted in the routine manner of CAS jaw reconstruction using FFF by the PA. Briefly, osteotomies, direct dental implant insertion, bone movements, and bone inset will be guided by the prepared surgical templates.
Intraoperative measurements will be conducted. Any FP or FN perforators will be recorded. Any unanticipated event that requires the modification of the fibula harvest plan will be recorded. Standard perioperative management will be conducted in a routine manner.
3. Postoperative evaluation Postoperative evaluation will be conducted between the surgical team led by the PA and the computer science team led by the Co-A. Any intraoperative events will be discussed in a timely manner. Data entry will be performed for each case right after the surgery.
(iii) Sample size The calculation of the sample size is determined by the primary endpoint, which focuses on the predictive accuracy of the AI tool in identifying the targeted skin perforators. As the success rate of the conventional method in identifying a reliable correlation was 67%, the aim of the AI-segmentation tool is to achieve a minimum accuracy of 80%. A one-sided test for non-inferiority/superiority is employed. The sample size is calculated using the exact binomial test for a one-sample non-inferiority/superiority clinical study. The hypothesis can be unified as: H0: p = p0 + δ versus Ha: p = p1. Here, δ represents the intervention effect size; p0 equals 80% and p is the true accuracy which is hypothesised to exceed 80% with a statistical power of at least 90% with a 5% level of significance.
According to a previous report on the AI-enabled skin perforator segmentation tool, the diagnostic accuracy was reported as 97%. The investigators can conservatively estimate the true success rate as 95% (representing an intervention effect size of 15%). Based on the calculation, a sample size of 44 is expected to have a 93.3% statistical power. Assuming at least one targeted perforator for each case and a drop-out rate after signing the informed consent of 10%, a total of 49 patients will be recruited to the study.
(iv) Data processing and analysis
All the data in this protocol will be recorded using a specially designed CRF (Appendix). Discrepancies in the database will be traced to the original documents for resolution. All records will be audited 100% for accuracy. Data entry will be performed by two independent assistants using the IBM SPSS Statistics Version 25. Consistency of input data between assistants will be compared using a inbuilt function.
For the demographic information, data will be presented as mean values with standard deviations for continuous data, and as counts with proportions for categorical data. The predictive accuracy will be calculated and presented as frequency. The one-sided 95% binomial proportion confidence interval will be computed and the lower confidence limit will be compared with 80%. The sensitivity, specificity, and positive and negative predictive values will be calculated. Continuous data (distance of perforators identified by the AI tool and during the surgeries, location of the perforators relative to the total fibula length, size of the perforators, time and satisfaction measurements, etc.) will be expressed as mean values with standard deviation (SD) for normally distributed data and expressed as the median with interquartile range (IQR) or range for skewed data. To evaluate the effect of various confounding variables on the precision of perforators identified by AI, suitable tests such as Pearson or Spearman's correlation test, t-Test, or Wilcoxon Signed-rank test will be selected depending on the normality of the data. Statistical significance will be set at p\< 0.05. All statistical analysis will be performed using IBM SPSS Statistics Version 25.
Recruitment & Eligibility
- Status
- NOT_YET_RECRUITING
- Sex
- All
- Target Recruitment
- 49
- Age ≥18 years, both genders;
- Provided signed and dated informed consent form;
- Indicated for immediate or secondary reconstructive surgery with osteocutaneous FFF.
- Patients who are pregnant;
- Patients who have medically compromised conditions and cannot tolerate surgery;
- Patients who are unable to receive pre-operative CT/CTA scans, such as those with iodine allergy;
- Patients who have anatomical variation preventing the safe harvest of FFF.
Study & Design
- Study Type
- OBSERVATIONAL
- Study Design
- Not specified
- Arm && Interventions
Group Intervention Description Patients requiring computer-assisted jaw reconstruction with microvascular free flaps AI Inclusion criteria 1. Age ≥18 years, both genders; 2. Provided signed and dated informed consent form; 3. Indicated for immediate or secondary reconstructive surgery with osteocutaneous FFF. Exclusion criteria 1. Patients who are pregnant; 2. Patients who have medically compromised conditions and cannot tolerate surgery; 3. Patients who are unable to receive pre-operative CT/CTA scans, such as those with iodine allergy; 4. Patients who have anatomical variation preventing the safe harvest of FFF;
- Primary Outcome Measures
Name Time Method predictive accuracy of the AI tool in identifying the targeted skin perforators 36 months The primary endpoint is the predictive accuracy of the AI-enabled skin perforator segmentation tool. When the skin perforator is identified by both the AI-segmentation tool and during the surgery, it will be counted as a true positive (TP). When the perforator is identified by the AI-segmentation tool, but not found during the surgery it is counted as a false positive (FP). When the perforator is seen during the surgery, but not shown by the AI tool, it is a false negative (FN). Finally, a true negative (TN) perforator count will be derived from those subjects, who do not exhibit an FN perforator. The predictive accuracy (PA) is identified as the percentage of true perforators among all the perforators, and is calculated as (TP +TN)/(TP+FP+TN+FN)\*100%.
- Secondary Outcome Measures
Name Time Method
Trial Locations
- Locations (1)
The Prince Philip Dental Hospital
🇭🇰Hong Kong, Hong Kong