Monday, April 3, 2023

Spontaneous regression of cancer


Spontaneous regression of cancer


Akseli Hemminki, MD, PhD, Professor of Oncology


Cancer Gene Therapy Group, Translational Immunology Research Program, University of Helsinki, Finland and Helsinki University Hospital Comprehensive Cancer Center, Finland.


Disappearance of tumors without treatment is a rare but well established phenomenon in medicine, and there are more than 4000 case reports found on this subject in PubMed. Once detected, cancer usually behaves quite predictably, progressing until successfully treated or until the patient dies. As the outcome is typically inevitable, those few cases that behave differently are noteworthy.

Putative explanations for spontaneous regression, also known as spontaneous remission, have ranged from divine intervention to diet, herbs, supplements and psychological factors. However, from the scientific point of view, the most likely explanation is symptomatic or asymptomatic infection, which leads to inflammation, stimulation of certain key receptors, and subsequent induction of an immune response against the tumor.

Indeed, many tumors are colonized with microbes.  On one hand, this demonstrates how the immune suppressed nature of tumors allows microbial growth. On the other, it could also indicate antitumor activities of microbes. The mechanisms of infection leading to immunity have been worked out in molecular detail, and proven in a randomized setting, while the same cannot be said for the other mechanisms proposed for spontaneous regression.

Estimates on the frequency of complete spontaneous regression vary wildly and there appears to be lack of reliable data. While numbers between 1:100 and 1:100 000 have been suggested, these could be under or overestimations as almost all reports are retrospective case studies. It is possible that some spontaneously regressing cancers were initially diagnosed incorrectly, and were in fact some other process. There is some thinking that partial or complete spontaneous regression might be especially common in certain “immunogenic” tumor types such as melanoma and renal cell cancer.

However, the clinical experience of most oncologists is that complete regression of advanced tumors, which have been confirmed by a pathologist, but not treated in any way, is exceedingly rare. Part of the reason could be that some treatment is usually attempted and thus response would be attributed to the attempt.

Inducing spontaneous regression

The earliest reports of engineering a spontaneous regression date back to Egyptian pharaoh Imhotep (2600 BC). In his time, inoperable tumors could be nicked with a knife, followed by application of a poultice on the wound.  No doubt many types of microbes were able to enter the tumor from the poultice.

St. Peregrine Laziosi constitutes a notable spontaneous regression case example from 1320. Divine intervention was proposed as the mechanism for a tumor first reddening, becoming inflamed and then disappearing. However, these findings would be quite compatible with a microbial infection.

Using bacteria against cancer

By the 18th century the association of microbial infection and tumor disappearance was becoming increasingly established. This led physicians to attempt purposeful infection of tumors and in 1813 Dr Vautier was experimenting with Clostridium perfringens as a treatment method. One notable scientist was Dr William B Coley, who in the late 19th century did ground-breaking work in optimizing “Coley’s toxin”, a mixture of bacterial supernatants. Coley’s toxin was ultimately shown to contain eg. tumor necrosis factor alpha and interleukin 12, cytokines which are actively studied for induction of anti-tumor immunity even today.

Use of wild type viruses in tumor therapy

Injection of bacteria into cancer patients tended to cause adverse events and therefore non-bacterial microbes became interesting. In 1896, tumor regressions were reported in “flu” patients, which led to experimental infection of cancer patients with various viruses in the early 20th century. While responses were reported, also side effects were common, as many of the viruses used were potent pathogens, including West Nile, Hepatitis, Influenza and others. Advanced cancer can cause immunosuppression, resulting in a higher likelihood of fulminant virus infection following injection of wild type viruses.

In the 1950s, adenovirus was discovered, and it was soon injected directly into cervical tumors. Responses were reported to occur in over half of the patients, and the treatment was well tolerated. Although viruses were now the most popular approach for creating “spontaneous regression”, also other microbes were still being studied. Bacillus Calmette Guerin, a weakened mycobacterium originally developed as a tuberculosis vaccine, was approved in 1977 for treatment of bladder cancer. This was also the first approval of an immunotherapy agent for tumor therapy.

Cancer immunotherapy

Purposefully engineered spontaneous regression can also be called cancer immunotherapy. Current approaches in immunotherapy include the use of viruses, cells, bacterial components, monoclonal antibodies, vaccines, and other approaches. With regard to viruses, progress in basic virology research ultimately allowed the generation of viruses modified so that they could no longer damage normal cells. These viruses are called oncolytic viruses, which essentially means a virus which has been altered so that it no longer replicates in normal cells, but is still able to replicate in tumor cells. The last step of such replication is lysis, breaking up of the tumor cell, which results in release of thousands of new virions into the surrounding tumor tissue and into the blood stream.

Oncolytic viruses

Dozens of different viruses have been studied as oncolytics. Early studies focused on unmodified wild type viruses, especially those that were less pathogenic than the viruses studied in the early 20th century. Newer “naturally occurring” oncolytic virus candidates include Newcastle disease virus, reovirus, coxsackievirus and a rat parvovirus. Many unmodified oncolytic viruses are originally from other species. Not being human pathogens, their ability to cause disease in humans is low. However, logically, also oncolytic potency appears lower with naturally occurring strains than with some of the most commonly used engineered human virus platforms, such as Herpes simplex, vaccinia and adenovirus.

The first oncolytic virus proven effective in a randomized trial was an adenovirus, H101 or Oncorine, which was approved in China in 2005. Patients with head and neck cancer were treated with chemotherapy, with or without Oncorine, the combination proving much more effective. The genome of H101 was engineered so that the virus can no longer bind and inactivate a key cellular growth control factor, p53, which is mutant in many tumors.

The next oncolytic virus approved for treatment of cancer was talimogene laherparepvec or Imlygic. Known also as T-Vec, this virus was studied in a randomized trial in melanoma patients. A key advantage over earlier constructs was an arming device, granulocyte macrophage colony stimulating factor (GMCSF), which had been engineered into the genome of the virus. This represents an important step in virus engineering; modifications were now being utilized to improve not only safety but also efficacy. T-Vec was approved in the US and Europe in 2015.

Oncolysis causes antitumor immunity

Infection with an oncolytic virus causes direct lysis of cancer cells, and this was initially thought to be the main anti-tumor mechanism. However, human data has shown that tumor-directed immune response, which results from oncolysis, may be even more important in overall efficacy. One reason for this is that tumors are, in fact, complex on the cellular level. They are not collections of identical cells. Instead, there are necrotic, hypoxic, hyperbaric and acidic areas, all of which present obstacles to progression of the oncolytic wave through the tumor. Also, tumors contain non-cancer cells such as fibroblasts, endothelial and immune cells, which together constitute the “stroma”. As oncolytic viruses only replicate in tumor cells, stromal barriers are a key limitation to direct oncolysis within tumors. Of important practical relevance, stromal barriers provide the rationale for repeated intratumoral injection, to allow more effective virus infection of different regions of the tumor.

Tumors that disappear without being detected

Presumably, most malignant cells are eradicated following recognition by the immune system. In nearly every case this happens on the microscopic level without the individual ever knowing that they had cancer. However, with improved cancer diagnostics, smaller and smaller tumors can now be detected. This has led to “overdiagnosis” of tumors that would not have affected the health of the individual during their lifetime. Organs such as the prostate, breast or thyroid may frequently have slow-growing or completely stagnant tumors which would not necessarily require removal. Some of these cases may represent immunological equilibrium, where immune response is able to control the tumor. Some of these tumors would perhaps eventually spontaneously regress or stay dormant until the individual dies of something else.                   

Nature’s own anti-cancer device?

Spontaneous regression is a known entity in oncology, but it is rare in most types of cancer. While oncolytic immunotherapy or checkpoint inhibitors do not work in every patient, responses are much more common than spontaneous regression. Why is this?

Probably the main reason is that when a cancer is diagnosed, the tumor already consists of hundreds of millions of cells, which have evolved under the strong selective pressure of the immune system over years or decades. In contrast, spontaneous regression is probably most common in subclinical situations, when tumor burden is minimal, and the malignant clones are not yet resistant to immune responses deriving from viral infection. Therefore, although challenging to prove, spontaneous regression may in fact be much more common that the frequency of cancer diagnosis. In contrast, it is difficult for the immune system, with or without viral infection, to overcome tumors which are large enough to be diagnosed.

One can speculate that viruses are nature’s own anti-cancer device. Humans have not gained full resistance to viral infections because they can perhaps be useful in situations where there are tumors which cannot be eradicated by the immune system alone. Viral infections typically occur through mucous membranes in the upper airways or the intestinal or urinary tracts and therefore viruses could be particularly useful in destroying tumors in these tissues. However, it is known that small amounts of virus can enter the bloodstream following local replication. This allows transduction of also distant tumors, for possible systemic antitumor effects. Some types of viruses are even able to hitchhike on cells of the blood to get to distant tumors.

The frequency of this phenomenon in humans is not known and it is difficult to study. Nevertheless, it is tantalizing to wonder if viral dormancy in the tonsils (adenovirus) or nerve ganglia (herpes) is allowed by the body for anti-tumor purposes? When tumors progress beyond a certain point, they typically induce immune suppression, which might then release viruses from these reservoirs into the blood, for destruction of the growing tumor?


In summary, clinical successes in immunotherapy have proven the notion that immunity resulting from viral infection can result in antitumor efficacy. This lends support to the notion that microbial infection can be a mechanism of “spontaneous regression”, and that in fact the immune system is probably able to eradicate most tumors without them becoming clinically evident.


I thank Joao M. Santos for help and comments. This blog is based on a book chapter cited as reference 1 below.

Conflict of interest

I am a shareholder in Targovax ASA, and employee and shareholder in TILT Biotherapeutics Ltd.


1.      Hemminki A, Santos JM. Oncolytic immunotherapy: from spontaneous regression to development of armed gene modified viruses.  The Evolution of Radionanotargeting towards Clinical Precision Oncology. Editor: Antti Jekunen. Bentham, 2022: 282-294. DOI: 10.2174/97816810886551220101

2.      Hemminki A. Crossing the Valley of Death with Advanced Cancer Therapy. Nomerta Publishing, Turku, Finland, 2015. ISBN: 978-952-7018-05-7. 230 pp. Paperback. ISBN 978-952-7018-06-4 (epub). Available at and