Multiple myeloma is the type of cancer that difficult to diagnosis early due to vague symptoms which can lead to misdiagnosis or create confusion with other diseases. When the symptoms are visible it may reach to advance cancer stage when even appropriate treatment is not left. This paper seeks and proposes a method for diagnosis of multiple myeloma through basic blood and urine tests by utilizing radar chart to increase readability and help to make rational decision making for patients and doctors. The radar chart is optimal way to visualizing numeric values. As a result of visualizing it as a graph, a clear visual and morphological difference was found compared to normal people.
Article
Diagnosis by visualizing blood data -multiple myeloma
https://doi.org/10.21203/rs.3.rs-2730376/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Blood cancer
Multiple myeloma
early diagnostic
classification
data visualization
decision making
Many scholars and medical scientists have long developed laboratory techniques for diagnosing various method to define diseases from a single drop of blood. This study began with the question of whether it is possible to interpret the numerical results of blood and urine tests in addition to laboratory techniques. In this paper, proposes visualizing method through the values derived from basic blood and urine tests to diagnosis Multiple Myeloma(MM). MM is a type of hematological malignancy that easily misdiagnosis and confused by other disease such as bone lesion, kidney disease and anemia or even consider it as part of aging. To diagnose MM, follow by blood chemistry test & urine tests first, and examine either one of the quantitative immunoglobulins test, Electrophoresis, Serum free light chains, and Bone marrow biopsy as well as bone X-rays, CT scans, MRI scans, PET scans, and more tests to definite diagnosis MM. However, there is a chance to find early through routine blood tests that shows an abnormally high amount of protein in the blood. Therefore, our study focus on the first step of diagnostic method of blood & urine test. Among the diagnostic criteria, features are selected only from possible results from blood and urine tests which are high blood calcium level(Ca), poor kidney function(Creatine), and low red blood cell (RBC(Red blood cell), Hemoglobin..) The research method steps are, collecting blood & urine test result data from target patients which are with/without symptoms and diagnosed with multiple myeloma, limited to before chemotherapy or under medicated treatment. Therefore, data were collected from 5 patients diagnosed MM, 1 patient without any blood related disease. After preprocessing the data, normalized it to be suitable for transformation into the graph. Among the many graph styles carefully selected radar charts that effectively show the status of blood test results inspired by the game character performance indicator graph and it is optimized for multivariate analysis and is useful for performance testing. The proposed method is expected to improve readability by visualizing the status of blood and urine test results, help medical staff make quick decisions, and help in early diagnosis.
The method used is to visualize the blood and urine test results, to explore data analysis (EDA). Through exploratory analysis, it summarizes the method of data normalization and establishes a formula. First, blood and urine test results collect data of 5 patients with multiple myeloma. One of the data is the blood and urine test result data collected four times after the symptoms are caused. And comparison data collects blood and urine data in women in their 30s who do not have blood -related diseases. Categorize the terminal patient and stage 1 patient among the collected patient data. Normalization of data is normalized to the same digit. Based on the normalized data, the graph is written on the radar chart, and the graph with the normal patient is compared and analyzed to make meaningful discovery.
Data & Variables
The blood test data of the patient comes from the open data source of ‘Korea Federation of Multiple Myeloma Patients’. Reference values are from Konkuk University medical center. The condition of the collected patient is as follows. Blood data was collected from 5 patients who is diagnosed with multiple myeloma without treated with drugs or chemotherapy. One patient is the results of 3 tests for three months of blood and urine tests that patient who died within eight months of diagnosis. One is a patient who died within three months of diagnosis, and the other 3 patients are the data of blood and urine tests diagnosed in the first stage of MM. (Table 1) The parameter value of each patient is white blood cell(WBC), neutrophil SEG%, lymphocytes, monocytes, eosinophils, basophil, red blood cells(RBC), hemoglobin(Hb), hematocrit(Hct), MCV (Mean corpuscular volume), MCHC (Mean corpuscular hemoglobin concentration), MCH (Mean corpuscular hemoglobin), albumin, and platelet(PLT). As well as Creatine and Calcium values, creatine indicates the condition of the kidneys, and calcium(Ca) indicates the broken bones, were added as parameters. (Table 1)
Table 1 Patient’s parameter value list & reference value
Table 2 Patient list and conditions
|
No. |
Patient list |
Conditions |
Symptoms |
|---|---|---|---|
|
1 |
Patient 1 |
first blood & urine tests after leg pain occur |
Y |
|
2 |
Patient 1–1 |
Patient 1's second blood & urine tests result after chest pain occur |
Y |
|
3 |
Patient 1–2 |
Patient 1's third blood & urine tests result: Decreased after 8 months after the tests |
Y |
|
4 |
Patient 2 |
Decreased: diagnosis after 3 months |
Y |
|
5 |
Patient 6 |
Diagnosis as Stage 1, No symptoms but failed to collect stem cell |
N |
|
Patient 7 |
Diagnosis as Stage 1, No symptoms |
N |
|
|
6 |
Patient 8 |
Diagnosis as Stage 1, Bone broke |
Y |
|
7 |
Patient 10 |
Normal, No blood related diseases |
Data Normalization
For accurate comparative analysis, it is necessary to normalize each patient’s parameter values of various parameters (1 to 100 units). The normalization of the data was applied according to the percentage method due to existence of international standard reference value for normal value of blood & urine. The conversion of percentage method is to visualize a graph round as possible for normal patient’s blood &urine test result. Also see what percentage of blood is composed of the maximum value based on the reference value. Explore patients' data before conversion, we find that most of patient’s values are lower than lowest reference value. distinctively at RBC parameter’s each patient’s value, 5 out of 7 patients are under lowest reference value, as well we Hct. Only Monocyte hit above highest reference value.(marketd in yellow) (Table 1) Therefore, normalizing data, we selected highest reference value to convert into a percentage. For the formula, P-value is each patient’s parameter value and R-value is reference value.
Formula:\(\text{X}=\frac{\text{P}-\text{v}\text{a}\text{l}\text{u}\text{e}}{\text{R}-\text{v}\text{a}\text{l}\text{u}\text{e}}\)
The method of analyzing the results is as follows. First, the difference between within 1 year terminal MM patients and normal patients is compared. Then compare the average value of the late hematological patients with normal patients (Fig. 1), a healthy woman in her early 30s. It is to compare some symptoms with a graph and normal patients and find a significance difference. The second compares symptom patients and normal patient’s graphs. The third expresses the graphs that obtain the average value of terminal cancer patients and the graphs of symptoms.
Explanation Of Test Results
Patient 1's data is a good example to see changes in blood and urine parameters over time. A total of 3 blood and urine tests were conducted for within 3 months, and the results are graphs applied as a conversion of percentage for the method of this study. (Figs. 1,2,3) The first graph shows the period in painkillers and rehabilitation treatment were used for lower leg bone pain. The second and third graphs show the times when patients complained of severe chest pain and back and leg pain. In the third graph where multiple myeloma was identified. Compared to the graph of a normal person, the white blood cell value of a normal person is 0.42 points, which is 0.22 points higher than the value of Patient 1 (0.64 point). Compared to normal patient, albumin and creatine, calcium segment showed in v shape of low creatine value. in patient 1 and 1–1,1–2 graphs noticeably increased in creatine and calcium levels over time. Other characteristics include Neutrophil Segmentaion %, lymphocytes, and lymphocytes segment area is rising like mountains in the case of monocyte sections. Through the 3rd examination, it can be seen that the dent in the opposite direction of the normal patient is more clearly visible and the depth gets deeper as time goes by. Another characteristic is that RBC decreases and WBC increases. In addition, remarkably low basophils(almost 0 value) and low PLT were observed, and albumin, creatine, and calcium fragments previously seen in patient 1 protruded in the opposite direction. (Fig. 1)
Patients who are diagnosed as stage I, have a similar appearance to those who are in the terminal stage. Three points can be identified. First, low RBC, low WBC, and low Hb intervals can be seen in the graph as being lower by 0.2 points compared to normal patients. The second is a high creatine fragment, showing the same pattern of patient 1 and patient 2. The third is a neutrophil, lymphocyte, and monocyte section that is not distinctly v-shaped, but is disproportionate and skewed to one side, and low basophils are also identified. (Fig. 2)
Finally, the average value of patients for each stage was obtained and the visualization method proposed in this study was applied. There is no significant difference between the shape of the graph of the terminally ill patient and the stage 1 patient.
This study was designed to diagnose multiple myeloma, a blood cancer that is difficult to diagnose, by reprocessing and visualizing primary data. Through visualization, the difference in graph form between normal patients and multiple myeloma patients was clearly found, and even when substituting asymptomatic patients, there was a big difference between normal patients, especially in the parameters of Platelet, Creaine, WBC, and Neutrophil segmentation % level. Therefore, not only saving many people’s lives, this study can provide opportunities for early diagnosis and treatment to patients who miss treatment opportunities by misdiagnosis, and is expected to contribute to readability and prompt or rational decision-making for medical staff by visually expressing numerical values. In addition, it is expected that false diagnoses can be reduced by preventing various diseases through early diagnosis or quickly determining the need for a second examination. As well as the visualization of blood is expected to further improve communication between patients and medical staff. Although this study was limited to a blood disease; multiple myeloma, it is expected to become a starting point for early diagnosis of various diseases by visualizing the state of blood and changing the parameter’s settings, which are diagnostic criteria for other cancer patients.
- Anderson KC, Alsina M, Atanackovic D, Biermann JS, Chandler JC, Costello C, et al. NCCN guidelines insights: multiple myeloma, version 3.2016. J Natl Compr Canc Netw 2016;14:389-400
- Korea Federation of Multiple Myeloma Patients, available at https://cafe.daum.net/overcomeMM
- Konkuk University Medical Center, available after login at https://www.kuh.ac.kr/main.do
There is NO Competing Interest.