Why Oncology Automation is a new ray of hope for cancer patients
Oncology Automation (OA) is bringing us to the root of cancer, helping us eliminate it without lengthy procedures and guesswork
In the days when ignorance was bliss, people died unaware they had cancer. Then, as humans evolved and got interested in the past, they began to dig up history and examine it. Cancer, it seems, was an old foe of humankind, and it was detected in mummies and fossils with a mention in ancient manuscripts.
Then came the age of diagnostics in medicine which brought cancer into our collective consciousness. We have been terrified of it since then. But with each passing decade, the fight against it is getting more potent as science and technology combine to lead to previously unachievable breakthroughs in treatment and cure.
This article sheds light on the advancements in cancer treatment powered by deep learning algorithms and biomedical procedures that help fight this scourge.
Oncology Automation Vs. Traditional Oncology Treatment
Before we talk of the advantages of oncology automation, let’s take a brief look at why treating cancer is a challenge.
When cells in the body turn rogue and begin to grow uncontrollably, spreading to other parts of the body, it’s like a senseless rebellion. We must kill the rogue cells to quell this anarchy without causing collateral damage, but our cellular matrix is a masterpiece of cohesion and integration. So, it makes it hugely challenging to excise only the cancer cells without affecting the neighbouring ones. Therefore, traditional cancer therapy affects cancerous and non-cancerous cells, diminishing the patient’s quality of life and lifespan, and leading to side effects. Thus, the biggest paradox oncologists were dealing with was the treatment standing in the way of the cure.
Life before Oncology Automation
Cancer treatment is a multidimensional challenge, and clinicians face numerous issues during the determination and treatment.
1) Prolonged Time and Human Errors:
During a biopsy, pathologists must repeatedly examine many slides for each patient to measure the cells’ length, surface area, and mitotic count. Often, they must apply additional immunohistochemical stains for quantitative analysis of the tissue section. This procedure makes the quantification process lengthy and may result in inefficacy and delayed treatment.
2) Difficulty in proper targeting of cancer cells
Radiation and chemotherapy are effective methods for cancer cell elimination, but they lead to collateral damage. Despite the radiologist’s judgment and dose administration, there is a high possibility that the normal, healthy cells of the body are also affected.
Chemo and radiation cannot judge the cancer cells’ borders; they do not follow the cancer-cell map with precision. Therefore, all the cells in the body receive the doses indiscriminately. As a result, when healthy cells are damaged, it has a domino effect and diminishes the patient’s tolerance for the next dose.
3) Overdosing and underdosing
In most cases of chemo and radiation, it becomes difficult for patients to tolerate the subsequent doses due to the above reasons. They fail to curtail the cancer growth effectively, which calls for a re-evaluation of the dose variation and dose difference depending on the cancer type and stage, i.e., early-stage or end-stage. The absence of such calibration means that the patient will either be overdosed or underdosed based on their tolerance levels.
Moreover, the second dose administration depends on the patient’s recovery from the effects of the first dose. Hence, the first dose should not exceed the dose tolerance limit of the individual. And here comes the crunch: when such dose calculations are done manually, the chances of underdosing or overdosing increase.
4) Inability to eradicate cancer cells
One of the significant threats to treatment is cancer cell recurrence. The cancer stem cells resist treatment and remain dormant, and this is a crucial concern because cancer stem cells have the potential to initiate tumour formation and relapse of cancer. Due to the high potential of cancer stem cells to bypass chemotherapy, the treatment cannot eradicate cancer. Also, there are cases where cancer cells have relocated and colonised some other part of the body, emigrating the disease bodily. It is difficult to free the body from all cancer cells through chemo or radiotherapy alone.
Enter Oncology Automation: Precision Targeting and No Collateral Damage
Automation is advantageous over traditional treatment in many ways:
- It reduces the detection time for cancer cells in a biopsy.
- Minimises human errors involved in the pathological quantification of cancer cells.
- Helps in the proper targeting of cancer cells during therapy.
- It helps in administering proper dosages.
- Destroys the source of the generation of cancer stem cells.
What is Oncology Automation (OA)?
It uses Artificial Intelligence to harmonise the multidimensional treatment of cancer. OA is a modern technology that reduces human intervention and the workload of clinicians. It is a computerised treatment plan that is precise and accurate, using less time to arrive at a diagnosis. It helps design a signature treatment for patients with aligned or similar pathological conditions using pre-recorded data.
Why is AI-led diagnosis and data analytics essential? The treatment and prognosis of cancer involve various challenges due to the properties of cancer cells, which are unlike normal cells. Cancer cells exhibit different behaviour:
- When a normal cell dies, it is a controlled aspect of the body’s regeneration. This process is called apotopsis, and signal receptors transmit the cell’s death. The absence of signal receptors means cancer cells do not die naturally in a cancer cell, and they survive longer and get more time to mutate, which can further tumour progression.
- The absence of signal receptors in cancer cells leads to their uncontrolled proliferation.
- Cancer cells also lack a protein called cadherin, which helps normal cells stay bound to each other. In the absence of this glue, cancer cells are free to move freely, which leads to them metastasising or increasing in number and invading the bloodstream.
Therefore, cancer cells are a proliferative, metastatic, and indefinite life span. As a result, they produce an abnormal colony of unwanted cells called tumours with no normal functions.
As a result of this wayward behaviour, cancer cells are a challenge to eradicate with precision. It’s like rooting out weeds in a vast, thick undergrowth without damaging healthy plants. It calls for precision in contouring the cancer-affected part, treatment planning, treatment delivery, accurate dosing, recording, and verifying the radiation delivered, aggregating data for analysis of radiation treatment, and quality assurance for the safety of the patients.
As we can see, it’s a big ask, and conventional cancer treatments often come up short on delivery. Oncology automation, on the other hand, effectively ticks these boxes.
The major areas where OA is replacing human intervention:
- Histopathological diagnostics by image scanning.
- Dose verification.
- Biomarker identification.
- Liquid biopsy.
- Radiogenomics.
Digital histopathological diagnostics using Deep Learning
Oncological Automation uses deep learning to improve the objectivity and efficiency of histopathology slide analysis. It has demonstrated this ability in prostate cancer identification in biopsy specimens and breast cancer metastasis in sentinel lymph nodes.
All slides of prostate cancer and macro and micro-metastases of breast cancer could be automatically identified, excluding the benign and normal cells, which make up 30 to 40 per cent of the slide. The Whole Slide Image (WSI) scanning system offers an opportunity to quantify and improve histopathologic procedures. The technology digitalises the biopsy glass slides with stained tissue sections at a high resolution, creating an invaluable reference archive for posterity.
Another advantage of digital WSI analysis is that it helps pathologists examine and quantify the slides without using multiple slides and other histochemical stains per patient, which vastly conserves time and effort.
Dose Verification in Radiotherapy by Deep Learning
Oncology Automation enhances treatment in the following areas:
- Planning and administration of radiotherapy.
- Auto contouring of typical structures of tumours (automatic tracing of tumour outlines in the body).
- Organ segmentation for treatment planning (Drawing the map of the body’s organs which need cancer treatment).
- Quality assurance and quality control procedures.
Dose verification, as mentioned earlier, is a critical aspect of cancer treatment. Adhering to radiotherapy quality assurance (QA) is time-consuming and laborious for physicists. Hence, a clinical mechanism to perform QA accurately for each patient is coming to light by UNet++ investigation to categorise the failure or success of Gamma-ray interaction.
Gamma passing rates (GPR) of more than 85% are considered passed, and less than 85% are considered failed. A highly complex process, UNet++ is a convolutional neural network (CNN) used for biomedical image segmentation. It is used to classify the passed or failed fields (whether the cancerous area received adequate radiation), predict dose differences, and has proved to be promising in quality assurance for radiotherapy.
Intensity-modulated radiotherapy (IMRT) is another method where the correct radiation dose to the cancer organs is calibrated and administered, ensuring the surrounding healthy tissues are unaffected.
Quality Assurance in oncology plays a critical role in confirming the accuracy of calculation of doses, the transmission of data, following up on linear accelerator performance, positioning of radiotherapy, and determining the accuracy of dosimeter response. QA compares the calculated dose distribution fluency with the measured dose distribution to ensure effective treatment administration maintains patient safety.
What are the latest radiotherapy procedures offered by Oncology Automation?
Proton therapy: Proton radiation is a more intelligent destroyer of cancer cells as compared to the use of photons. It leads to minor damage to healthy tissues. Photon radiation targets the tumours with precision using a minimal dose. It reduces overall toxicity by minimising short and long-term side effects of radiotherapy. It scores over other radiotherapeutic agents in treating recurrent tumours.
Cyberknife: It is a non-invasive surgical procedure for inoperable, complex tumours. It can treat tumours from more than 1,400 angles without harming other tissues.
Robotic surgery: Robotic surgery is a robot-assisted surgery that works with more accuracy, flexibility, and control done through a minimum incision.
True beam STX: TrueBeam STX is an advanced linear accelerator – another gift of Oncology Automation that targets the hard-to-reach organs. It is the fastest in treatment, with more accuracy and simplicity.
What are the other advanced biomedical cancer treatments?
Knowledge of cancer’s origins at the genetic level is crucial to defeating it.
The study and alteration of the mutated gene of the tumour DNA is the only way to permanently rid of it.
The steps towards knowing cancer better are:
Collection of circulating tumour DNA by liquid biopsy.
The conventional cancer detection method has been collecting the suspect tissue of the affected organ and subjecting it to a biopsy.
Liquid biopsy
Biopsy of the tissue sections of tumour cells is an invasive procedure, and today’s advancements allow cancer detection through non-invasive means such as blood, saliva, urine, cerebrospinal fluid etc. Genomic examination of circulating tumour DNA (ctDNA) from these sources helps clinicians detect, diagnose, and monitor cancer at an early stage. It can also track patients’ responses to treatment or surveillance for those who have already completed treatment but are at a high risk of recurrence.
There is evidence of liquid biopsy tests’ ability to track the treatment response in patients with lymphoma, which showed the positive changes in ctDNA. It also provides additional information regarding drug resistance of the cancer cells, which is a result of specific mutations. There is a record that patients with lung cancer develop resistance to the treatment by Tyrosine Kinase Inhibitors (TKI) within 1-2 years of administering the drug. TKI is an oral drug.
A significant factor in detecting cancer cells is that not all tumour cells are cancerous. They are often heterogeneous by nature which indicates not every cell’s DNA undergoes mutational changes. Only the cancerous cells undergo mutation to initiate the disease.
So, we must study two types of mutations to understand the cause of cancer. One is called driver mutation, and the other is passenger mutation. As the name suggests, driver mutation creates the conditions for cancer, promoting the rogue growth of cells. On the other hand, passive mutations do not actively drive cancer development but can be present in cells where driver mutations exist.
Biomarker identification
Biomarkers are molecules of DNA, RNA, ctDNA (circulating tumour DNA), proteins, or exosomes that show alterations in cells that signal cancer. They are easily identified as the derivatives of cancer cells.
The genetic alterations in the DNA of the tumour cells are solely responsible for the emergence of characteristics of cancer cells, and these can be detected in miRNA.
Why is micro-RNA or miRNA the key to cancer cure?
The micro-RNA or miRNA are small, non-coding, endogenous RNAs that regulate many critical cellular functions, including multiplication, differentiation, and death of cells. They also hand in the invasion, growth, and generation of cancer stem cells (CSC).
Suppose these functions can be directly identified and the genetic alterations in ctDNA rectified. In that case, it can lead to the breakthrough we have been waiting for – eliminating cancer at its causal level.
If we can deregulate the role of miRNA, it will lead to the arrest of invasion, proliferation, survival, metastasis, and drug resistance of the cancer cells. Hence miRNA based therapeutic approaches are the go-to pursuits in cancer treatment.
Radiogenomics using Artificial Intelligence: The Pay-off
Once the cancer biomarkers are identified, radiogenomics help with their thorough examination to understand the cause of cancer.
Radiogenomics comprises the genomic and proteomic (analysing the protein component) study of cancer cells to understand their cause and mutation. Once the conversion is understood, it can help do away with the cause of the formation of cancer cells.
Oncology Automation takes the image analysis of radiogenomics a step further by looking at imaging phenotype. It studies the physical form and structure of the organism, how it is developing, its biochemical and physiological properties, behaviour, etc.
This examination and analysis play a significant role in understanding cancer’s profile, and the data can also be used to correlate with other data by comparison or contrast.
Conclusion
We know that cancer treatment poses multiple challenges for oncologists. With the advent of Oncology Automation, the prospects for cancer treatment have brightened significantly. Analysing and computing the clinical parameters before starting treatment is the way to go, and its advantages are clear.
The value of human intelligence lies in its ability to acknowledge its limitations. The birth of technology is a result of this awareness. As we design more effective technology, it confirms our need to combine the might of human endeavour and human-designed machine power to bring us the solutions we need.
The advent of Oncology Automation offers Precision, Perfection, and a Pathway in cancer treatment without invasive methods, lengthy procedures, endemic organic obstacles, and collateral damage.
