Cancer Detection by means of Mechanical Palpation

Transcription

1 Cancer Detection by means of Mechanical Palpation Design Team Paige Burke, Robert Eley Spencer Heyl, Margaret McGuire, Alan Radcliffe Design Advisor Prof. Kai Tak Wan Sponsor Massachusetts General Hospital Abstract Early detection of breast cancer has been shown to increase patient survival rates; therefore an early detection system that identifies lesions under the skin could help save lives. Potentially cancerous tissue masses have a higher modulus of elasticity making them stiffer than the surrounding healthy tissue. A subsurface sensing device is needed because all current methods of detection are either expensive or require a trained technician to read the results. The goal of this device is to detect an inconsistent tissue mass as small as 5 mm in diameter. This 5mm size lesion is classified as Stage I breast cancer. For the scope of this project, the detection is accomplished by simulating the movement of a cancerous mass in human tissue, which is represented by a ball bearing and a phantom gel, respectively. When the mechanical device contacts the tissue surrounding the lesion, it detects a change in the force caused by an increase in total tissue stiffness. The group designed and prototyped the sensor head as well as the data collection method. Test Frame For more information, please contact k.wan@neu.edu. Sensor Head Assembly 135

2 The Need for Project Current breast cancer detection methods lead to delays in diagnoses due to physician bias and lack of training. Currently, mammography and clinical breast exams (CBE) are the most commonly used breast cancer detection methods. Mammography uses non ionizing radiation to image the breast tissue, but tumors can be missed due to size or location in the tissue. A CBE involves a physician palpating the tissue with three fingers to systematically search for lesions. Both of these methods incorporate physician bias which can lead to delay in diagnoses and a worse prognosis for the patient. Misread results and physician bias stem from lack of experience and training. There is also a discrepancy in the tactile sensitivity of physicians performing the clinical breast exams which can lead to a misdiagnosis. Creating a device that can eliminate the subjectivity of these current methods can help find a cancerous mass earlier and improve the prognosis of the patient. The Design Project Objectives and Requirements The objective of this project is to create a device that uses palpation theory to detect a 5 mm ball bearing in a silicone phantom at a maximum depth of 30 mm. Design Objectives The goal of this project is to create a device that uses the mechanical properties of the breast to detect lesions. This device must detect and locate lesions more accurately and consistently than an annual CBE. In palpation theory, the pop is the rapid movement of a lesion from compressed tissue to uncompressed tissue. The device will detect and locate lesions by capturing changes in force that come from the pop of a lesion within the tissue. In order to read the force created by the movement of the lesion within the tissue, the design must compress the tissue at a consistent depth and measure the change in force. Design Requirements The device must create minimal friction as well as create the pop by palpating a silicone phantom with a sensor head. It must recognize and locate the pop, measuring force values and coordinates as well as having a force accuracy of +/- 0.2 N. It must detect lesions as small as 5 mm, at depths up to 30 mm. 136

3 Design Concepts Considered Three concept designs were produced and evaluated using a decision matrix. The initial concepts included rollers and segmented indenters coupled with force sensors. Figure 19: Sensor Head Option 1 Figure 20: Sensor Head Option 2 Figure 21: Sensor Head Option 3 The main preliminary designs focused on the same general concept. This concept involved an indenter head, with one or more force sensors attached, indenting the tissue in order to capture a force reading. The drastic change in stiffness as the indenter head is moved along a silicone phantom and comes into contact with a simulated lesion will cause the sensors to read a higher force than the force of the silicone alone. The figures shown to the left are the three main designs that incorporate this concept. In Figure 1, the red roller indents the skin and makes contact with the protruding edge of the force gauge (shown in brown). The force values would be continually read and saved as the device moved around the examination area. In Figure 2, the large rounded edge of the device indents the tissue and the roller, shown in red, has force sensors laid around it. The idea behind this roller design is to create a more comfortable experience for the patient. Again, the sensors will detect a change in force when a lesion is present and store that data for further evaluation. In Figure 3, the indenter is split up into nine pieces (shown in red) that are all connected to a transfer plate (shown in black). The indenter pieces are able to move vertically, but only slightly. In between the transfer plate and the indenter pieces are sensors that will recognize the change in force as the indenter pieces move across the gel. This allows nine force readings to be taken at once. The force data collected from these designs will be read into an excel file. From there, the excel data is be uploaded to and analyzed using MATLAB code written by the group. This code analyzes each force reading from each sensor and keeps track of the x and y location of the forces that are significantly larger than the average of force of the phantom alone. The user will be presented with a screen that shows the x and y locations of potential lesions and a 3D graph of the x & y positions and the force (shown in Figure 6). The group used a decision matrix to objectively rank each design and in the end, sensor head option 3 (Figure 3) was chosen as the final design. However, several modifications were made to this design to increase its potential. 137

4 Pins 5 Individual Pieces Transfer plate Tekscan Force Sensors Recommended Design Concept The final sensor head design breaks a half sphere into five individually actuated parts, allowing for five independent Figure 22: Final Sensor Head Design Figure 23: Final Sensor Head Design Figure 24: Graphical Output of force data from MATLAB code force readings Design Description The final design, shown in Figure 4, incorporates the same concept as the sensor head option 3 (depicted in Figure 3) but has a semi-sphere split into 5 different pieces. This gives the device a rounded point of contact and allows it to collect force data and move easily in both the x and y directions. The pins (seen in gray) are actuated against force sensors. Each section of the sensor head can move independently from the others allowing for five individual force readings. This allows for better tracking of the lesion against the indenter. The Tekscan force sensors chosen for this application are extremely thin and small, allowing for flexibility in the design. Larger force sensors would have not allowed for as many sensors or would have required a heavy final design. The final sensor head assembly (shown in Figure 5) has the force sensors mounted to the top green transfer plate. The test frame used for this project motorizes the motion of the sensor head in order to move it in the x and y directions along a horizontal plane. The test frame also provides a firm platform for the silicone phantom to be indented upon. Experimental Investigations Preliminary testing with an Instron machine indenting a silicone phantom provided a value of approximately 40 N of force to indent up to 30 mm. A value of 60 N was used for all design calculations in order to give a factor of safety of at least 1.5. The final design uses rounded edges to prevent the silicone from ripping at forces that may reach beyond 60 N. The height of the sensor head is longer than the indentation depth of 30 mm to ensure the silicone only comes into contact with the sensor head. Key Advantages The main advantage of this design is that it objectively collects and analyzes the force data. In comparison to the current methods of detection, which incorporate physician bias, this can help prevent misdiagnoses of cancerous masses. Another key advantage of this design is that it is made for comfort. The entire procedure will use a maximum of 40 N to push the indenter head into the tissue. While 40N is a significant amount force, it is only one quarter of the amount of force applied to the breast tissue during a mammogram. 138

5 Financial Issues With an initial budget of Working with an initial budget of $700.00, the main costs of this $700.00, this project is prototype are found in the sensors, motors and linear carriages. Most expected to cost a total of mechanical components such as pulleys and belts cost less than $20.00 $ each. The total price of this project, including the test frame, is estimated to reach $ Mass manufacturing costs are estimated to be $ per unit or less, assuming bulk purchases of at least one thousand parts. It should be noted that this prototype is designed for bench top testing only, and has not yet been optimized for real world applications. Recommended Improvements Adding motorized control of the vertical motion and additional force sensors to measure lateral forces will add to the accuracy of data collection for the future. While the x and y motion of the sensor head is motor controlled, the motion of the base plate is controlled manually. A small scissor lift provides reasonable indentation control, but a motorized system would allow for the sensor head to indent to a more exact depth over a multitude of tests. Another improvement for the base plate is angular rotation capabilities. If the angle of attack is changed, the potential of detecting a cancerous lesion increases. Another improvement involves the sensor head. The current design only measures the vertical component of detectable forces. Adding force sensors to capture horizontal forces is a recommended addition to the sensor head. Testing has indicated that while the vertical forces are most important, there is valuable information that can be gained through horizontal force measurements. 139