Cancer During Pregnancy in the UK: What the CARING Study Found

Table of Contents

• Key Points
• Background: Why This Research Matters
• How the Study Was Done
• Who Was Included: Patient Characteristics
• How Cancers Were Investigated and Staged
• Treatments Given During Pregnancy
• Outcomes for Mothers and Babies
• How These Results Compare With International Data
• What This Means for Patients
• Limitations: What the Study Could Not Prove
• Recommendations and Next Steps
• Frequently Asked Questions
• Source Information

Key Points

• In a UK study of 119 women diagnosed with cancer during pregnancy, 80.7% were planned for curative treatment and 81.7% of pregnancies ended in a live birth.
• In that study, 20.2% of women died during the study period; gastrointestinal cancers had the highest death rate at 45.5% (5 of 11 patients).
• Just over half of the women (57.1%) started cancer treatment during pregnancy, most often surgery (37.0%) or chemotherapy (26.1%); no patient received radiotherapy.

Background: Why This Research Matters

More women are being diagnosed with cancer during pregnancy. Researchers point to two main reasons. First, women in developed nations such as the UK are choosing to have children later in life. Second, the risk of most cancers rises with age.

Counting these cases is harder than it sounds. Not every pregnancy ends in a live birth, and health registries do not routinely combine cancer records with maternity records. Despite these obstacles, researchers estimate that cancer affects between 17 per 100,000 live births and 25 to 27 per 100,000 pregnancies. That works out to about two new cases every day in the UK.

Newer screening tests add another layer. Non-invasive prenatal testing (a blood test that looks for chromosomal abnormalities in the baby) uses cell-free DNA from the mother's blood. Doctors have reported cases where this test led to a cancer diagnosis in women who had no symptoms at all.

Pregnancy complicates every step of cancer care. Doctors must weigh the mother's need for treatment against the safety of the unborn child. Parents and physicians also face difficult decisions about both cancer care and maternity care. This study aimed to fill a UK data gap. The researchers set out to answer four questions:

1. How many cases of cancer during pregnancy occurred in the UK over five years, and what types, stages, recurrence rates and timing were involved?
2. What did the women's demographic and maternal profiles look like?
3. What treatments did they receive, by tumour type and by stage of pregnancy?
4. What happened to mothers and babies immediately after birth?

UK guidance from the Royal College of Obstetricians and Gynaecologists sets out general rules for treatment. Surgery can usually take place in any trimester (three-month stage of pregnancy), but it may not be feasible depending on where the tumour sits. Radiotherapy is usually contraindicated (not advised) but may be used in specific situations, such as preserving life or function, with shielding to protect the baby. Systemic chemotherapy (drugs that travel through the bloodstream) should be avoided in the first trimester. Doctors consider it safe from the second trimester onwards. Even so, birth should not happen more than 2 to 3 weeks after the last chemotherapy dose. This gap lets the bone marrow recover, so the risk of neutropenia (dangerously low white blood cell counts) falls. Immunotherapies, hormone therapies and targeted therapies are contraindicated until after delivery.

After a live birth, women receiving chemotherapy, immunotherapy, hormone therapy or targeted therapy should not breastfeed. Women having radiotherapy may need to avoid close contact with their newborn and other young children. Finally, because a cancer diagnosis in pregnancy is frightening and complex, women and their families often need extra psychological and practical support.

How the Study Was Done

This was a retrospective study (a study that looks back at existing medical records). Researchers included women diagnosed with cancer during pregnancy between 1 January 2016 and 31 December 2020. The National Oncology Trainee Collaborative for Healthcare Research (NOTCH) network carried out the data collection.

Fourteen cancer centres took part. Ten were in England, three in Scotland and one in Wales. Together they served an estimated population of 21 million people. Sites identified patients in different ways, most commonly through clinician recall (doctors remembering cases) or electronic coding.

The team designed a data collection template using local systems at each site. They mapped the data fields to the European International Network on Cancer Infertility and Pregnancy (INCIP) database. They chose fields that were feasible to collect within UK National Health Service (NHS) oncology and obstetrics frameworks.

Data collection focused on three areas: how the cancer first showed up during pregnancy, the cancer treatment that followed, and the outcomes for mother and baby immediately after birth. Long-term infant follow-up was not included.

The primary investigator collated data from all sites and cleaned and anonymised it (removed identifying details). An independent analyst then ran the statistics. To protect patient privacy, the team grouped rarer tumour types together whenever fewer than five patients shared a category. All analyses used RStudio version 4.3.2. Local Caldicott Guardian approval (the UK framework for protecting patient confidentiality) was obtained at every site. Because the study was retrospective, patient consent was not required.

Who Was Included: Patient Characteristics

Sites submitted data for 144 patients. Of these, 119 were eligible for analysis. Sixteen were excluded because their diagnosis fell outside the study period, and nine were excluded because their diagnosis was ineligible or did not occur during pregnancy. Eighty-eight patients came from the five centres that used electronic records. Cases were spread evenly across the five years, with an average of 24 cases per year across all sites.

The median age at diagnosis was 33 years (range 22 to 49). The most common primary sites (where the cancer started) were:

• Breast — 30 patients (25.2%)
• Skin — 25 patients (20.0%)
• Haematological (blood cancers) — 18 patients (15.1%)
• Genitourinary (ovary, cervix and urinary tract) — 14 patients (11.8%)
• Gastrointestinal (gut, upper and lower tract) — 11 patients (9.2%)
• Other sites — 21 patients (17.7%), including ear, nose and throat, lung, thyroid, neurological and unknown primary cancers

Most cancers were new diagnoses — 109 patients, or 91.6%. Ten patients (8.4%) had a recurrence (the cancer came back) of a previous cancer. Four of these recurrences started in the breast, and three of those four were oestrogen receptor positive (meaning the cancer grows in response to the hormone oestrogen). This finding highlights the need for close monitoring during pregnancy, especially in women with a previous history of cancer.

Most women (82.4%, 98 patients) had symptoms when they were diagnosed. Only 21 patients (17.7%) had no symptoms. Symptom patterns differed sharply by cancer type. More than half of women with genitourinary cancers were diagnosed without symptoms (51.1% in the study text; 8 of 14 patients, or 57.1%, in the tumour-by-tumour breakdown). Symptomatic presentation was far more common for other sites, ranging from 80% for skin cancer to 100% for gastrointestinal cancer.

The women were mostly of any white ethnic background (91 patients, 76.5%), with 13 patients (10.9%) from non-white backgrounds and ethnicity missing for 15 patients (12.6%). About half had never smoked (59 patients, 49.6%). Seven patients (5.9%) were current smokers, 18 (15.1%) were ex-smokers, and smoking status was unknown for 35 patients (29.4%).

The median body mass index (BMI) was 24.9 kg/m². The median gravidity (total number of pregnancies) was 2, with a range of 0 to 8. The median parity (number of births) was 1, with a range of 0 to 4. Just over a third of women (34.2%) were nulliparous (had never given birth). Most pregnancies followed natural conception (99 patients, 83.2%). Seven patients (5.9%) used assisted conception, and conception method was unknown for 13 patients (10.9%).

Cancer stage (how far the cancer had spread) was known for most women. Stage I accounted for 34 patients (33.0%), stage II for 18 (17.5%), stage III for 9 (8.7%) and stage IV for 20 (19.4%). Stage was not available for 22 patients (21.4%).

Timing of diagnosis varied across pregnancy. Sixteen women (13.5%) were diagnosed in the first trimester, 46 (38.6%) in the second trimester and 37 (31.1%) in the third trimester. The trimester was unknown for 20 women (16.8%).

How Cancers Were Investigated and Staged

Staging tests matter because they determine what treatment is possible. Just over half of the patients (56.7%) had more than one staging investigation. About a quarter had no staging investigation recorded during pregnancy — 28 patients, or 23.3% to 23.5% depending on the breakdown.

Skin cancer dominated this group. Twenty of those 28 patients (71.4%) had skin cancer. Eleven had basal cell carcinoma (a slow-growing skin cancer that rarely spreads) and nine had malignant melanoma (a more serious skin cancer).

Doctors used several imaging methods during staging:

• Ultrasound — 55 patients (46.2% to 46.7%)
• MRI or diffusion-weighted MRI (magnetic resonance imaging, which does not use radiation) — 60 patients (50.4%)
• X-ray — 21 patients (17.7%)
• Mammography (breast X-ray) — 19 patients (16.0%)
• CT scan (computed tomography, which uses ionising radiation) — 16 patients (13.5%)
• Bone scan — 11 patients (9.2%)
• Endoscopy (a camera test inside the body) — 8 patients (6.7%)

No patients had a PET-CT scan (a combined imaging test). The number of investigations per patient varied: 28 patients (23.5%) had none, 24 (20.2%) had one, 30 (25.2%) had two, 25 (21.1%) had three and 12 (10.1%) had four or more.

The researchers noted an important safety point. Most imaging used non-ionising methods, which do not expose the baby to radiation. However, 16 women (13.5%) had a CT scan and 11 (9.2%) had a bone scan. Doctors do not generally recommend either of these tests in pregnancy because of the risk of radiation exposure to the baby. The authors call for more research into whether these scans followed specific safety guidance, and whether the benefits outweighed the risks.

Treatments Given During Pregnancy

Most patients were diagnosed at an early stage and planned for curative treatment. Of the whole group, 96 patients (80.7%) received radical treatment (treatment given with the aim of curing the cancer). Twenty patients (16.8%) received palliative treatment (treatment to control the disease and relieve symptoms rather than cure it). Treatment intent was unknown for three patients (2.5%).

Just over half of the women — 68 patients, or 57.1% — started some form of cancer treatment during pregnancy. The rest did not. The specific treatments were:

• Surgery — 44 patients (37.0%)
• Chemotherapy — 31 patients (26.1%)
• Radiotherapy — no patients
• Other therapies (including targeted therapy, immunotherapy, hormone therapy and steroids) — 10 patients (8.4%)

Counting all treatments, 51 patients (42.9%) had no treatment during pregnancy, 54 (45.4%) had one treatment and 14 (11.7%) had two or more.

Timing followed the general safety rules. Of the 37 patients with surgical timing recorded, 5 (13.5%) had surgery in the first trimester, 17 (46.0%) in the second and 15 (40.5%) in the third. Just under half of all operations — 21 of 44, or 47.7% — were abdominal or cervical surgery. Of the 25 patients with chemotherapy timing recorded, 17 (68.0%) were treated in the second trimester and 8 (32.0%) in the third. No patient received chemotherapy in the first trimester.

Chemotherapy drugs varied. Nineteen patients (16.0%) received alkylating agents other than platinums, 18 (15.1%) received taxanes, 14 (11.8%) received anthracyclines, 9 (7.6%) received antimetabolites, 8 (6.7%) received platinum drugs and 7 (5.9%) received other agents.

How much treatment a woman received depended heavily on when she was diagnosed. Among women diagnosed in the third trimester, 56.4% had no active cancer treatment during pregnancy. The figure was 35.9% for those diagnosed in the second trimester and just 2.2% for those diagnosed in the first trimester.

Treatment intent also varied by tumour type. Gastrointestinal cancers stood out. Seven of 11 patients with gastrointestinal cancer (63.6%) were treated with palliative intent, and only 4 (36.4%) with radical intent. By contrast, all 16 patients with haematological cancers with known intent (100%) were treated with radical intent, along with 27 of 30 breast cancer patients (90%), 22 of 24 skin cancer patients (91.7%) and 13 of 14 genitourinary cancer patients (92.9%).

Outcomes for Mothers and Babies

Birth outcomes were generally good, but not without loss. There were 98 live births (81.7%), 10 terminations of pregnancy (8.4%, reported as 8.3% in the abstract), 7 miscarriages (5.9%; reported as 8, or 6.7%, in the results text) and 4 other outcomes (3.4%).

Of the live births, 52 babies (53.1%) were delivered by caesarean section and 46 (47.0%) by vaginal delivery. The abstract reports a slightly higher caesarean figure of 54 (55.1%). The caesarean rate varied by cancer type:

• Genitourinary cancers — 90.0%
• Gastrointestinal cancers — 75.0%
• Haematological cancers — 58.8%
• Other cancers — 52.9%
• Breast cancer — 42.3%
• Skin cancer — 40.9%

Preterm birth (delivery before 37 weeks of pregnancy) was common. Among live births, 34 babies (42.5%) were estimated to be preterm. Of these preterm births, 27 (79%) were delivered by caesarean section. Across the whole group, 35 deliveries (29.4%) were preterm, 47 (39.5%) were full term, and the term was unknown for 37 (31.1%). Full-term birth rates also differed by tumour type, ranging from 5.0% for skin cancer to 75.0% for genitourinary cancer, with breast cancer at 52.6%, haematological cancer at 53.3% and gastrointestinal cancer at 66.7%.

Live birth rates by tumour type were similar for most groups: 26 of 30 breast cancer patients (89.6%), 22 of 25 skin cancer patients (88.0%), 16 of 18 haematological cancer patients (89.0%), 18 of 22 patients with other cancers (90.0%). The rates were lower for gastrointestinal cancer (7 of 11, 63.6%) and genitourinary cancer (9 of 14, 64.3%).

Maternal outcomes were the most sobering part of the study. At the point when each site stopped collecting data, 24 patients (20.0% to 20.2%) had died from disease progression. Of the 95 women still alive, 75 (78.9%) were in remission (no sign of cancer), 12 (12.6%) were still on treatment and 8 (8.4%) were lost to follow-up. Looking at the whole group, 75 patients (63.0%) were in remission, 12 (10.1%) had persistent or recurrent disease and 24 (20.2%) had died.

Death rates varied widely by cancer type:

• Gastrointestinal cancers — 45.5% (5 of 11 patients)
• Other cancer sites — 38.1% (8 of 22)
• Breast cancer — 20.0% (6 of 30)
• Haematological cancers — 11.1% (2 of 18)
• Skin cancer — 8.0% (2 of 25)
• Genitourinary cancers — 7.1% (1 of 14)

The authors suggest the high gastrointestinal death rate reflects later-stage diagnosis at presentation, since 63.6% of these patients were treated with palliative intent.

How These Results Compare With International Data

The median age of 33 years matches previous studies, including a report of 1,170 cases from the INCIP group. The ethnic mix matched UK census data and national maternity audits, which suggests the dataset broadly represents the UK population. Compared with the INCIP group, fewer UK women were first-time mothers — 34.2% here versus 44% to 56% internationally. This matters because previous research suggests whether a woman already has children can influence decisions about continuing a pregnancy or delaying treatment.

Breast, skin and haematological cancers were the most common in this UK cohort, which fits the published literature. However, a National Cancer Registry Audit covering 2012 to 2015 found cervical cancer to be the third most prevalent, at 15.2% of all cases. This study captured many other primary sites too, showing the wide range of cancers that can affect pregnancy.

Treatment patterns differed from international data in one key way. Only 26.0% of patients here received chemotherapy during pregnancy, compared with 37% reported by INCIP. The authors say this may point to underuse of chemotherapy in the UK and call for further investigation. Surgical patterns were more comparable: 21 of 44 operations (47.7%) were abdominal or cervical, similar to INCIP findings.

Mortality was higher than international figures. One in five women in this cohort died. Even among women with breast cancer — the most studied group — crude mortality was 20% here versus 14% in the INCIP study. The authors caution that these comparisons are difficult because most prognostic studies focus on single tumour types.

The authors also compared their findings with the MBRRACE maternal mortality report, which covers UK women who died during pregnancy or within 42 days of pregnancy ending. For 2019 to 2020, that report found breast, ovary and cervix cancers were indirect causes of death, at a rate of 0.19 per 100,000 maternities. Gastrointestinal cancers had the highest mortality rate in this study, which may reflect more aggressive tumour biology or later presentation.

The authors describe a phenomenon called "diagnostic overshadowing." Symptoms such as fatigue, abdominal pain or gastroesophageal reflux (heartburn) can be wrongly attributed to normal pregnancy, delaying a cancer diagnosis. For context, Kocian and colleagues previously reported that 41.5% of pregnant women with gastrointestinal cancers had stage 4 disease.

What This Means for Patients

This is the first pan-tumour report — meaning it covers many cancer types at once — of diagnosis, management and outcomes for cancer diagnosed during pregnancy in the UK. It proves that collecting this data is feasible. It also shows how much more research is needed.

Several messages stand out for patients and families:

• Most cancers in this group were treatable. Four in five women (80.7%) were planned for curative treatment.
• Most pregnancies continued to a live birth. Nearly 82% of pregnancies ended in a live birth.
• Treatment during pregnancy is possible and common. Over half of women (57.1%) started treatment while pregnant, most often surgery or chemotherapy.
• Timing of diagnosis shapes treatment. Women diagnosed later in pregnancy were far less likely to receive active treatment before delivery.
• Gut cancers behaved worst. They were more often advanced at diagnosis, more often treated with palliative intent, and had the highest death rate.
• Outcomes were worse than international averages. The 20% death rate and the lower chemotherapy use warrant further study.

The authors argue strongly for a national approach to linking obstetric (pregnancy) data with cancer data. Pregnancy is a protected characteristic under the UK Equality Act 2010. In practice, that means the NHS should be able to audit treatments and outcomes for these women and their families. The NHS should use standardised datasets at least as good as those collected internationally.

Limitations: What the Study Could Not Prove

The authors are candid about the study's weaknesses. The NOTCH network is made up mostly of oncologists who treat solid organ cancers. As a result, patients with haematological (blood) cancers may be under-represented compared with the real population.

Because this was a preliminary study, investigators could not access the children's medical notes. That means no infant outcomes beyond the immediate postnatal period were available. International datasets such as INCIP include these longer-term outcomes, so this UK study is less complete in that respect.

The study was retrospective, so it relied on records rather than planned data collection. Some data points, such as ethnicity, had higher rates of missing information. Diagnosing timing was unknown for 16.8% of patients, and staging was unavailable for 21.4%.

Comparing outcomes across countries is also hard. Most prognostic studies focus on one tumour type, such as breast cancer, rather than a mixed group like this one. Finally, the study could not draw firm conclusions about how treatment choices directly affected maternal outcomes.

The authors do highlight real strengths. Because the study was designed around readily available data points, very little information was missing. Completeness was comparable to maternity services data and better than studies using routine cancer data. Oncologists entered the data by hand after reviewing case notes, which supports its reliability.

Recommendations and Next Steps

The authors call for several actions based on these findings:

1. Build a national linked dataset. Combine obstetric and cancer records across the UK so every case can be tracked and audited.
2. Investigate chemotherapy use. Only 26.0% of patients received chemotherapy during pregnancy versus 37% internationally. Researchers should examine whether UK patients are missing out on safe treatment.
3. Study diagnostic pathways. More work is needed to understand how pregnant women reach a cancer diagnosis and what barriers or enablers affect how quickly they present.
4. Examine radiation safety. Future research should check whether CT and bone scans in pregnancy followed safety guidance and whether benefits outweighed risks.
5. Improve awareness of diagnostic overshadowing. Symptoms such as fatigue, abdominal pain and reflux should not be dismissed as pregnancy alone when they persist or are unusual.
6. Provide psychosocial support. Women and families facing cancer in pregnancy need enhanced psychological and supportive care.

If you are pregnant and notice symptoms that are new, persistent or unusual for you, speak to your midwife or doctor. Ask for them to be investigated rather than assumed to be part of pregnancy. If you have a history of cancer, tell your maternity team early, because this study found that women with previous cancer can experience recurrence during pregnancy.

Frequently Asked Questions

How common is cancer diagnosed during pregnancy?

Researchers estimate cancer affects between 17 per 100,000 live births and 25 to 27 per 100,000 pregnancies. That works out to about two new cases every day in the UK. Counting cases is difficult because not every pregnancy ends in a live birth, and health registries do not routinely combine cancer records with maternity records.

Can cancer be treated during pregnancy?

Yes. In a UK study of 119 women, 57.1% started some form of cancer treatment during pregnancy, most often surgery or chemotherapy. UK guidance says surgery can usually take place in any trimester. Chemotherapy should be avoided in the first trimester but is considered safe from the second trimester onwards. Radiotherapy is usually not advised.

What are the risks to the baby if I have cancer treatment while pregnant?

In the UK study of 119 women, 81.7% of pregnancies ended in a live birth. Preterm birth was common: among live births, 34 babies (42.5%) were estimated to be preterm. UK guidance says chemotherapy should stop 2 to 3 weeks before birth so the bone marrow can recover and the risk of low white blood cell counts falls.

Which cancers diagnosed in pregnancy have the highest death rate?

In the UK study of 119 women, gastrointestinal (gut) cancers had the highest death rate at 45.5% (5 of 11 patients). The authors suggest this reflects later-stage diagnosis, since 63.6% of these patients were treated with palliative intent. Death rates were 20.0% for breast cancer, 11.1% for blood cancers, 8.0% for skin cancer and 7.1% for genitourinary cancers.

What is diagnostic overshadowing and why does it matter in pregnancy?

Diagnostic overshadowing is when symptoms such as fatigue, abdominal pain or heartburn are wrongly attributed to normal pregnancy, delaying a cancer diagnosis. In the UK study of 119 women, gastrointestinal cancers had the highest death rate at 45.5%. If you are pregnant and notice symptoms that are new, persistent or unusual for you, ask your midwife or doctor to investigate them.

I had cancer before. What should I tell my maternity team?

Tell your maternity team early if you have a history of cancer. In the UK study of 119 women, 10 patients (8.4%) had a recurrence of a previous cancer during pregnancy. Four of these recurrences started in the breast, and three of those four were oestrogen receptor positive. This finding highlights the need for close monitoring during pregnancy.

When should a patient diagnosed with cancer during pregnancy seek a second opinion?

A second opinion can help when the tumour type, stage or treatment plan is unclear, since 21.4% of women had no stage recorded and 23.5% had no staging investigation during pregnancy. It is also worth seeking when gastrointestinal cancer is diagnosed, as 63.6% of these patients received palliative rather than curative treatment and 45.5% died. Because only 26.0% of women received chemotherapy during pregnancy, compared with 37% internationally, reviewing whether safe treatment was missed is reasonable. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: CAnceR IN PreGnancy (CARING) – a retrospective study of cancer diagnosed during pregnancy in the United Kingdom

Authors: M.A. Baxter, M. Denholm, S.J. Kingdon, S. Kathirgamakarthigeyan, S. Parikh, R. Shakir, R. Johnson, H. Martin, M. Walton, W. Yao, A. Swan, C. Samuelson, X. Ren, A. Cooper, H-L Gray, S. Clifton, J. Ball, G. Gullick, M. Anderson, L. Dodd, H. Hayhurst, M. Salama, R. Shotton, F. Britton, T. Christodoulou, A. Abdul-Hamid, A. Eichholz, R.M. Evans, P. Wallroth, F. Gibson, K. Poole, M. Rowe, J. Harris, and the National Oncology Trainees Collaborative for Healthcare Research (NOTCH).

Journal: British Journal of Cancer (2024), volume 130, pages 1261–1268

Publication details: Received 24 November 2023; revised 25 January 2024; accepted 29 January 2024; published online 21 February 2024. DOI: 10.1038/s41416-024-02605-x. Open access, distributed under a Creative Commons Attribution 4.0 International License.

Corresponding author: m.z.baxter@dundee.ac.uk (Division of Molecular and Clinical Medicine, Ninewells Hospital and Medical School, University of Dundee, Dundee, UK)

This patient-friendly article is based on peer-reviewed research. It describes a retrospective study of 119 patients and does not replace individual medical advice. Please discuss any concerns about cancer in pregnancy with your own clinical team.