A review of biopsies carried out at the Bristol Royal Infirmary (BRI) indicates that almost 7,000 serious errors could have been made over the course of a decade.
The inquiry was carried out after doctors raised concerns with managers about the standards of pathology at the hospital, following a number of misdiagnoses.
Cases include a former nurse who is now terminally ill after lymphoma was missed; a former NHS manager of cancer services who died when breast cancer went undetected; a patient who died from invasive skin cancer, which was missed; and a woman who had part of her breast removed because cancer was wrongly diagnosed.
The investigation, ordered by the United Hospitals Bristol Trust (UHBT) which runs the infirmary, examined 26 specific allegations of misdiagnosis, which had been raised by the whistle-blowers, and found that just seven of the diagnoses were correct.
Buried in an annexe to the inquiry's report was an admission that an audit of a further 3,500 cases found that in 3.4 per cent of them, independent pathologists who re-examined the samples believed the wrong diagnosis had been made.
With the hospital examining more than 20,000 specimens a year, the figures could represent up to 6,800 blunders over a decade.
In a report submitted to the inquiry, Prof Peter Furness, then President of the Royal College of Pathologists called for all the cases involving discrepancies to be re-reviewed, in order to establish definitive error rates – which require three opinions. Yet that was never done.
Instead the inquiry, paid for by the hospital, said it had found no evidence the overall service was not safe, when it reported in December 2010.
No action was taken against any pathologists about whom concerns had been raised.
Now, patient campaigners have written an open letter to the chairman of the inquiry, exposing the figures. They fear that plans to merge the BRI pathology unit with services across the city could put a larger population at risk.
The group, led by Daphne Havercroft, a former breast cancer patient who helped to bring whistle-blowers' concerns to public attention, say the proposals are "reckless and dangerous" because the problems at the unit have not been properly investigated.
Between 2004 and 2009, at least 10 NHS doctors, including six consultants, most working at nearby North Bristol trust, brought concerns about the safety of BRI's department to the attention of their own managers and those at the BRI.
When no action was taken, their correspondence, detailing a series of alleged blunders and fears about wider failings, was made public, triggering the inquiry.
One of the doctors, Dr Nassif Ibrahim, a consultant pathologist at North Bristol trust, who retired last year, said the investigation had been "a whitewash from the start".
"A three per cent error rate would never be acceptable," he added. "I would not have wanted a biopsy from a member of my family to go there."
Evidence to the inquiry told how one patient was diagnosed with a fatal tumour, on the basis of a biopsy carried out at BRI and gave away £30,000 thinking that he did not have long to live.
When the case was reviewed it was found that the man in fact had a rare but treatable chronic infection.
None of the patients affected by the misdiagnoses, or those bereaved as a result, were told about the 18-month inquiry into the allegations until days before the report was published.
Catherine Calland, a former nurse who is now terminally ill, after cancer was missed. She had been initially told that a lump found in her breast during a routine mammogram in 2005 was benign. A year later she was informed that it was a malignant lymphoma.
It was removed but despite undergoing radiotherapy Mrs Calland's condition deteriorated. The lymphoma, which was at its earliest stage when it was missed, had become too advanced to cure.
A consultant denied there had been any mistake in the original biopsy analysis. The grandmother of two was only notified that an error had been made when she received a letter, four days before the inquiry report was published.
Mrs Calland said doctors must have realised the error when they looked at the results of her second mammogram. Pathologists told her it was routine procedure to re-examine the results of a previous biopsy when a malignancy was found.
She said: "Why was I lied to for so long when I asked if there had been a mistake in my original diagnosis?"
"I find the behaviour of certain individuals in this case appalling," she said. "Patients should to be at the very heart of everything the NHS does. To lie to patients and to protect those who make mistakes is simply atrocious."
Jane Hopes, a former assistant director of cancer services at North Bristol trust died from breast cancer in 2004, after a lump was wrongly classed as benign.
She was never told of the misdiagnosis, nor was her widower Reg Hill. He was not told that the case was part of the inquiry, until The Sunday Telegraph contacted him.
Mr Hill said: "If they are going to operate under the same system, with the same people, you are simply going to get more of these cases of misdiagnosis in the future."
Mrs Calland and Mr Hill are now preparing to sue the hospital, in cases taken on by Laurence Vick, who led the families' legal team at the 2001 public inquiry into the deaths of more than 30 babies at the same hospital.
Mr Vick, from legal firm Michelmores, said he was concerned that the hospital had "learned nothing" since that inquiry.
He accused the hospital trust of "an alarming lack of candour" in its response to patients who had been misdiagnosed and in failing to respond to whistle-blowers' concerns.
UHBT said it had apologised to patients harmed, accepted the inquiry's findings and focused on implementing the recommendations.
A spokesman said the inquiry's independent panel found no evidence to suggest that the department was not safe.
An inquiry spokesman said "The panel did not re-review [the slides in which discrepancies were found] because the panel felt it would not materially affect their conclusions."
He said there was no clinical consensus about what constitutes an acceptable error rate.
